{ "version": 1, "updated_at": "2026-10-20", "intents": [ { "intent_id": "tech-path", "user_question_examples": [ "Why does my website load but AI does not recommend it?", "If Google found an article, does that mean ChatGPT already knows the company?", "What does the “submitted” status mean in AURA GEO?", "Can Auri show exactly what has already been done for the website?" ], "article_path": "/en/learn/kak-sajt-popadaet-v-otvety-ii/", "title": "How a website appears in AI answers: five distinct stages", "excerpt": "First a page must be accessible, then discovered and selected for a specific answer. A citation means a link to a source; a recommendation means suggesting that someone choose a company or product. These are different outcomes, and each requires its own evidence.", "updated_at": "2026-10-20", "capability": "url_evidence_pipeline", "tool_names": [ "aura_get_onboarding", "aura_get_url_status", "aura_get_citations" ], "prerequisites": [ "A website or public URL", "For project data, sign-in and access to the project" ], "expected_result": "You know which stage is confirmed and what to check next.", "limitations": [ "Accessibility does not guarantee indexing or a recommendation.", "The external agent must actually obtain the service’s answer." ] }, { "intent_id": "work-fit", "user_question_examples": [ "What is AURA GEO and why does a small business need it?", "Does Auri promote my website in every AI by itself?", "I need a service for my website and my own AI agent—what should I look for?", "Can AURA be used as just another chat?" ], "article_path": "/en/learn/kogda-podhodit-aura-geo-i-chto-delaet-auri/", "title": "When AURA GEO helps a website and what role Auri plays", "excerpt": "AURA GEO fits when you need to prepare a website for AI search step by step and see the results of completed actions. It brings together audits, tasks, content and observations. Auri helps navigate the process; your connected agent performs work on files and AI services.", "updated_at": "2026-10-20", "capability": "An overall website-preparation workflow and saved results", "tool_names": [ "aura_get_onboarding", "aura_get_agent_brief", "aura_get_recommendations", "aura_get_client_report" ], "prerequisites": [ "A public website for practical work", "A compatible agent to perform tasks" ], "expected_result": "A clear work sequence and sources verifying the result", "limitations": [ "Does not guarantee citation", "Does not replace every web analytics tool", "A separate Auri API model is not claimed" ] }, { "intent_id": "measure-buyer-questions", "user_question_examples": [ "How can I find out which questions my buyers ask AI?", "Should I add the company name to every prompt?", "Which questions should I add to AURA GEO first?", "Can Auri discover the queries of every ChatGPT customer?" ], "article_path": "/en/learn/voprosy-pokupatelej-dlya-monitoringa-ii/", "title": "Which buyer questions to test in AI: unbranded and branded", "excerpt": "To check whether a company is discovered, start with questions about choosing a solution without naming it. Keep branded questions separate: they show what the assistant says about a company already named. These represent different buyer intentions and different grounds for deciding which content to improve.", "updated_at": "2026-10-20", "capability": "plan", "tool_names": [ "aura_get_questions", "aura_create_questions", "aura_update_question" ], "prerequisites": [ "Public material is available without signing in. Project data requires authorized access; writes are performed only within granted permissions." ], "expected_result": "A verifiable conclusion and next step tied to an actual question, answer or event, with no guarantee of indexing or citation.", "limitations": [ "A project's sample does not represent all AI users' questions and answers." ] }, { "intent_id": "measure-signals", "user_question_examples": [ "If ChatGPT names the company, is that a recommendation?", "Why does AURA separate links from mentions?", "What does an unknown recommendation mean?", "There is an AI link: how many customers arrived?" ], "article_path": "/en/learn/upominanie-citirovanie-rekomendaciya-ii/", "title": "AI mentions, citations and recommendations: three different results", "excerpt": "A mention is a brand name or domain in the text. A citation is a link to a website page used in the answer. A recommendation is an explicitly positive context for choosing a product. None of these signals alone proves a visitor referral or purchase. Read the entire answer first, then count the relevant metric.", "updated_at": "2026-10-20", "capability": "citations", "tool_names": [ "aura_get_observations", "aura_get_citations", "aura_get_ai_traffic" ], "prerequisites": [ "Public material is available without signing in. Project data requires authorized access; writes are performed only within granted permissions." ], "expected_result": "A verifiable conclusion and next step tied to an actual question, answer or event, with no guarantee of indexing or citation.", "limitations": [ "A project's sample does not represent all AI users' questions and answers." ] }, { "intent_id": "work-connect-agent", "user_question_examples": [ "How do I connect AURA GEO to Codex and confirm everything works?", "I added the MCP URL; why can the agent not see anything?", "Do I need a separate plugin for Auri to work with my website?", "Can I connect AURA to Claude Code without AI-provider keys?" ], "article_path": "/en/learn/podklyuchit-mcp-agenta-i-proverit-instrumenty/", "title": "How to connect your own agent to AURA GEO through MCP", "excerpt": "The existing AURA GEO MCP server address is https://aurageo.io/mcp. Add it to a compatible client, complete sign-in and test access with aura_list_projects. A saved configuration without a successful server response does not yet mean the connection works.", "updated_at": "2026-10-20", "capability": "Connect the existing remote MCP server", "tool_names": [ "aura_list_projects", "aura_get_agent_brief", "aura_get_onboarding" ], "prerequisites": [ "A compatible MCP client", "An AURA account and completed authorization" ], "expected_result": "The client reads an authorized project and its current next step", "limitations": [ "Interface support depends on the client", "MCP does not provide sessions for external AI services", "Plugin visibility and identifiers remain unchanged" ] }, { "intent_id": "work-first-site", "user_question_examples": [ "Can I simply give Auri a link instead of filling in every field?", "How do I add a client website through MCP if it is not in the workspace yet?", "Why will AURA not let the agent open an existing website?", "Will a www page link create a separate project?" ], "article_path": "/en/learn/dobavit-sajt-cherez-agenta-po-ssylke/", "title": "How to add a website to AURA GEO through an agent with one link", "excerpt": "You can give a connected agent a website address and ask it to create a project. It needs configuration permission and authorization to create projects. After creation, it reads onboarding, obtains the next step and continues while retaining information already known.", "updated_at": "2026-10-20", "capability": "Project creation by URL and managed onboarding", "tool_names": [ "aura_list_projects", "aura_create_project", "aura_get_onboarding", "aura_update_project" ], "prerequisites": [ "MCP authorization", "write_project and permission to create projects" ], "expected_result": "One project with an exact domain and an available next action", "limitations": [ "A local address is unsuitable for external crawling", "Creating a project does not verify ownership", "Do not duplicate an inaccessible existing project" ] }, { "intent_id": "tech-answer-first", "user_question_examples": [ "What text should I add to my website for AI citations?", "Do I need a separate page for every question?", "Can Auri help prepare an article brief?", "Do three numbers and a table guarantee a ChatGPT link?" ], "article_path": "/en/learn/stranica-otvet-s-faktami-i-istochnikami/", "title": "How to write an answer page that is easy to verify and cite", "excerpt": "Start with a useful answer to one reader need. Then explain the conditions, steps, an example and evidence. A concise opening helps readers find their way, but a fixed length or special phrase does not guarantee a citation.", "updated_at": "2026-10-20", "capability": "evidence_backed_content", "tool_names": [ "aura_get_answer_workspace", "aura_upsert_answer_page", "aura_publish_answer_page" ], "prerequisites": [ "One reader need", "Genuine facts and sources", "Publication permission and a verified publishing location" ], "expected_result": "A useful standalone article with verifiable claims.", "limitations": [ "Answer-first is an editorial approach, not a guaranteed position.", "Generated text requires fact-checking." ] }, { "intent_id": "measure-evidence-modes", "user_question_examples": [ "Does reopening an AURA report ask ChatGPT again?", "Can Codex check a topic instead of Perplexity?", "What is saved alongside an answer?", "Has Auri already checked every AI in the background?" ], "article_path": "/en/learn/sohranennyj-otvet-novyj-zapros-i-issledovanie-agenta/", "title": "Saved answers, new AI requests and agent research", "excerpt": "Reading history returns a previously saved observation. A new response is produced after an actual request to an accessible AI. Research by Codex or another agent is a separate mode: it must not be presented as an answer from consumer ChatGPT, Gemini or Perplexity. Every result should include its mode, source and time.", "updated_at": "2026-10-20", "capability": "citations", "tool_names": [ "aura_get_observations", "aura_get_monitoring_run", "aura_claim_monitoring_check", "aura_save_observation" ], "prerequisites": [ "Public material is available without signing in. Project data requires authorized access; writes are performed only within granted permissions." ], "expected_result": "A verifiable conclusion and next step tied to an actual question, answer or event, with no guarantee of indexing or citation.", "limitations": [ "A project's sample does not represent all AI users' questions and answers.", "AURA checks the consistency of saved metadata but does not authenticate an answer's origin with the AI provider; the evidence remains a report from the connected agent." ] }, { "intent_id": "tech-business-identity", "user_question_examples": [ "How should I describe a company if AI confuses it with a brand of the same name?", "What should I tell Auri about the company before a GEO audit?", "Which links should I add to sameAs for a new service?", "Does AURA GEO need fake reviews or a Wikidata item to get cited?" ], "article_path": "/en/learn/opisanie-kompanii-i-brenda-dlya-ai-poiska/", "title": "How to describe your company and services so AI does not confuse your brand with namesakes", "excerpt": "Start with a consistent description of the real business: one primary name, the official domain, a clear category, services and a working contact. Use this information on the homepage, About page and in structured data. If the name is common, clarify which organization you mean; do not replace unknown facts with invented ones.", "updated_at": "2026-10-20", "capability": "Business identity context and a task to update public information", "tool_names": [ "aura_get_onboarding", "aura_update_project", "aura_get_generated_files", "aura_create_geo_tasks" ], "prerequisites": [ "Genuine information from the owner", "Permission to change the project through MCP", "Separate website access for public edits" ], "expected_result": "A consistent business description in the project and a verifiable plan to correct public pages", "limitations": [ "AURA context is not an independent company registration check", "Filling in sameAs does not guarantee citations", "External profiles and reviews must not be invented" ] }, { "intent_id": "tech-content-plan", "user_question_examples": [ "How do I choose ten useful articles from a hundred customer questions?", "Can Auri turn a FAQ into a GEO content plan?", "Should I write a separate article for every query formulation?", "How can I use competitors’ sources in AURA GEO for a content plan?" ], "article_path": "/en/learn/plan-materialov-iz-voprosov-pokupatelej/", "title": "How to turn customer questions into a content plan for AI search", "excerpt": "A plan starts with customer needs, not the number of articles to produce. Collect questions, combine those with the same intent and check where the website already answers them. Then decide whether to improve an existing page, publish a standalone piece or first gather missing facts. Separately save neutral questions for subsequent monitoring.", "updated_at": "2026-10-20", "capability": "Content planning based on intentions and recommendations from saved sources", "tool_names": [ "aura_get_questions", "aura_get_observations", "aura_get_recommendations", "aura_get_geo_tasks", "aura_create_geo_tasks", "aura_get_answer_workspace" ], "prerequisites": [ "Audience questions with identified origins", "A review of existing pages", "Verified facts for the selected topics" ], "expected_result": "An intent map with primary URLs, a focused priority plan and completion criteria", "limitations": [ "Agent suggestions are not AI query volumes", "A source gap requires reading the original page", "Producing content does not prove indexing or citations" ] }, { "intent_id": "measure-faq-vs-monitoring", "user_question_examples": [ "Why do we need questions if the website already has an FAQ?", "Why is an FAQ excluded from an AURA GEO run?", "Can answers be retained after changing a question's purpose?", "How can Auri help turn an FAQ into a plan?" ], "article_path": "/en/learn/faq-i-monitoring-voprosov-v-chem-raznica/", "title": "Why a website FAQ cannot automatically count as demand monitoring", "excerpt": "An FAQ helps visitors understand a known product. Monitoring checks which answer AI gives to a neutral buyer question. The same wording can sometimes serve both purposes, but its role and expected result must be defined in advance. The number of FAQ pages is not the number of customer queries checked.", "updated_at": "2026-10-20", "capability": "plan", "tool_names": [ "aura_get_questions", "aura_update_question", "aura_get_monitoring_run" ], "prerequisites": [ "Public material is available without signing in. Project data requires authorized access; writes are performed only within granted permissions." ], "expected_result": "A verifiable conclusion and next step tied to an actual question, answer or event, with no guarantee of indexing or citation.", "limitations": [ "A project's sample does not represent all AI users' questions and answers." ] }, { "intent_id": "measure-comparable-runs", "user_question_examples": [ "Why do two reports show different visibility?", "What does questionSetHash mean in AURA GEO?", "Can a temporary chat be compared with a regular one?", "How should blocked checks in the second run be handled?" ], "article_path": "/en/learn/kak-sravnivat-progony-monitoringa-ii/", "title": "How to compare AI checks and distinguish changed results from changed conditions", "excerpt": "Compare answers to identical questions under conditions that are as similar as possible. Changing the language, model, panel or search mode changes what is being measured. Two attractive summary figures do not by themselves constitute a valid trend: first check the question set, collection method and completeness of each run.", "updated_at": "2026-10-20", "capability": "citations", "tool_names": [ "aura_get_monitoring_run", "aura_begin_monitoring_run", "aura_get_observations" ], "prerequisites": [ "The article is publicly readable. Observations require a selected project and authorized access; new checks require an AI interface that is actually accessible." ], "expected_result": "A comparable, verifiable conclusion from the saved sample with a specific next action.", "limitations": [ "One panel's result is not the volume of all queries and does not guarantee a future AI answer." ] }, { "intent_id": "measure-small-sample", "user_question_examples": [ "Why does ChatGPT change its service list?", "Do two links out of four mean 50% of the market?", "Can AURA prove the probability of a recommendation?", "How many checks are enough for a guarantee?" ], "article_path": "/en/learn/pochemu-otvety-ii-menyayutsya-i-kak-chitat-maluyu-vyborku/", "title": "AI linked to the site today but not tomorrow: interpreting a small sample", "excerpt": "One saved answer confirms only an observation under specific conditions. Assessing consistency requires repetitions and a clear sample size. Show the number of checked answers alongside the share, keep unsuccessful and blocked attempts separate, and do not select only favorable texts for the report.", "updated_at": "2026-10-20", "capability": "citations", "tool_names": [ "aura_get_monitoring_run", "aura_get_observations" ], "prerequisites": [ "The article is publicly readable. Observations require a selected project and authorized access; new checks require an AI interface that is actually accessible." ], "expected_result": "A comparable, verifiable conclusion from the saved sample with a specific next action.", "limitations": [ "One panel's result is not the volume of all queries and does not guarantee a future AI answer." ] }, { "intent_id": "measure-competitors", "user_question_examples": [ "Why does AI recommend a competitor but not me?", "How should competitors be chosen for AURA GEO?", "Does mention count indicate market share?", "Can I copy the article AI cites?" ], "article_path": "/en/learn/sravnenie-konkurentov-v-otvetah-ii/", "title": "How to compare competitors in AI answers without an invented ranking", "excerpt": "Compare competitors within the same panel and conditions: who is mentioned, who is explicitly recommended and whose pages are linked. The number of saved panel answers is not market share, brand popularity or proof of purchases. A useful result is an explainable gap and a suitable action.", "updated_at": "2026-10-20", "capability": "plan", "tool_names": [ "aura_add_competitors", "aura_get_observations", "aura_get_recommendations" ], "prerequisites": [ "The article is publicly readable. Observations require a selected project and authorized access; new checks require an AI interface that is actually accessible." ], "expected_result": "A comparable, verifiable conclusion from the saved sample with a specific next action.", "limitations": [ "One panel's result is not the volume of all queries and does not guarantee a future AI answer." ] }, { "intent_id": "measure-source-gaps", "user_question_examples": [ "How can I identify useful places to discuss the company?", "Does AURA detect the competitor on every cited website?", "What is a source-gap task?", "Why does a source list not bring new customers?" ], "article_path": "/en/learn/istochniki-ii-i-zadachi-po-probelam/", "title": "How to turn AI citation sources into an improvement plan", "excerpt": "Start with a source in a saved answer: which question does it help explain, and what is your page missing? A nearby competitor mention does not prove that the competitor appears on the external page. Read the source first, then decide whether you need your own content, a factual clarification or a suitable external publication.", "updated_at": "2026-10-20", "capability": "plan", "tool_names": [ "aura_get_observations", "aura_get_recommendations", "aura_get_answer_workspace", "aura_list_actions" ], "prerequisites": [ "The article is publicly readable. Observations require a selected project and authorized access; new checks require an AI interface that is actually accessible." ], "expected_result": "A comparable, verifiable conclusion from the saved sample with a specific next action.", "limitations": [ "One panel's result is not the volume of all queries and does not guarantee a future AI answer." ] }, { "intent_id": "measure-referral-traffic", "user_question_examples": [ "How many people came to my website from ChatGPT?", "Why are some AI referrals recorded as direct?", "Can an AI link be tagged for AURA GEO?", "Does a referral reveal the question the customer asked?" ], "article_path": "/en/learn/kak-schitat-perehody-iz-ii-referrer-utm/", "title": "How to measure AI referrals and account for missing sources", "excerpt": "An actual referral is measured by a load event on the connected website, not by the presence of a link in an answer. AURA recognizes some AI sources from a referrer or supported tag. If no source is transmitted, it cannot reliably be reconstructed from total visits and attributed to a specific assistant.", "updated_at": "2026-10-20", "capability": "analytics", "tool_names": [ "aura_get_site_installation", "aura_configure_analytics", "aura_get_ai_traffic", "aura_get_insights" ], "prerequisites": [ "The article is public without registration. Your own analytics require a selected project, verified website and active connections for the relevant sources." ], "expected_result": "An understanding of the measurement unit, available coverage and a specific action for obtaining actual data.", "limitations": [ "AURA does not reconstruct unknown sources or count unique buyers from tracker events." ] }, { "intent_id": "measure-conversions", "user_question_examples": [ "How can I tell whether an AI visitor submitted an inquiry?", "Does AURA count a ‘Buy’ button as a sale?", "What is the difference between client_conversion and server_conversion?", "Can Auri show every future deal from one visitor?" ], "article_path": "/en/learn/perehody-zayavki-i-konversii-iz-ii/", "title": "How to connect AI referrals with inquiries without counting a click as a sale", "excerpt": "A pageview is not an inquiry, and a button click does not guarantee a saved lead or payment. Define the required event, configure its delivery and distinguish a browser message from backend confirmation. AURA counts these categories separately; they do not automatically become unique buyers.", "updated_at": "2026-10-20", "capability": "analytics", "tool_names": [ "aura_get_site_installation", "aura_configure_analytics", "aura_get_insights", "aura_get_ai_traffic" ], "prerequisites": [ "The article is public without registration. Your own analytics require a selected project, verified website and active connections for the relevant sources." ], "expected_result": "An understanding of the measurement unit, available coverage and a specific action for obtaining actual data.", "limitations": [ "AURA does not reconstruct unknown sources or count unique buyers from tracker events." ] }, { "intent_id": "measure-crawler-proof", "user_question_examples": [ "Does OAI-SearchBot in the log mean ChatGPT is citing me?", "Why does AURA call a bot request unverified?", "My audit got 200 with Googlebot: is the website indexed?", "Do I need to enable GPTBot for ChatGPT Search?" ], "article_path": "/en/learn/vizit-ii-bota-ne-raven-citirovaniyu/", "title": "A bot visited the site: does that mean indexing and citation?", "excerpt": "A record bearing a bot's name confirms a received report about a request. A User-Agent can be reproduced, so provider ownership is checked separately. Even an authenticated crawl does not prove search inclusion, use in an answer or a buyer referral. Each further conclusion needs its own evidence.", "updated_at": "2026-10-20", "capability": "analytics", "tool_names": [ "aura_get_bot_visits", "aura_get_url_status", "aura_verify_urls", "aura_get_job" ], "prerequisites": [ "The article is available without signing in. Working with private data requires an authorized project; execution requires current permissions and accessible external tools." ], "expected_result": "A verifiable action plan and an explanation of the result with saved evidence and limitations.", "limitations": [ "Connection, a technical check and a saved report do not guarantee an AI recommendation or a customer outcome." ] }, { "intent_id": "measure-empty-history", "user_question_examples": [ "Why is there no answer history in AURA GEO?", "If there is no data, does AI not see my website?", "Can past referrals be obtained from a website link?", "What should Auri say when there are no observations?" ], "article_path": "/en/learn/net-dannyh-v-monitoringe-ii-s-chego-nachat/", "title": "No answers or referrals in history yet: what to do first", "excerpt": "No saved observations means AURA currently has no project history to show. It is not proof that the company is absent from every AI. Start with one suitable question and an actual answer; connect referral analytics separately. Do not fill an empty screen with demonstration figures presented as the client's results.", "updated_at": "2026-10-20", "capability": "citations", "tool_names": [ "aura_get_questions", "aura_get_observations", "aura_get_monitoring_run", "aura_get_site_installation" ], "prerequisites": [ "The article is available without signing in. Working with private data requires an authorized project; execution requires current permissions and accessible external tools." ], "expected_result": "A verifiable action plan and an explanation of the result with saved evidence and limitations.", "limitations": [ "Connection, a technical check and a saved report do not guarantee an AI recommendation or a customer outcome." ] }, { "intent_id": "measure-access-blocked", "user_question_examples": [ "Why can the agent not see my AURA project?", "Does signing in to AURA with Google automatically connect Gemini?", "Does a CAPTCHA mean the website is not recommended?", "Can Auri continue without publication permission?" ], "article_path": "/en/learn/net-avtorizacii-ili-proverka-ii-zablokirovana/", "title": "No authorization or an AI check is blocked: how to continue", "excerpt": "Public instructions require no sign-in. Private AURA history requires project access, while a new check of a particular AI needs an accessible interface for that service. If any level is missing, state the exact reason and one necessary owner action; do not bypass authorization or invent a result.", "updated_at": "2026-10-20", "capability": "citations", "tool_names": [ "aura_list_projects", "aura_get_monitoring_run", "aura_finish_monitoring_check", "aura_get_observations" ], "prerequisites": [ "The article is available without signing in. Working with private data requires an authorized project; execution requires current permissions and accessible external tools." ], "expected_result": "A verifiable action plan and an explanation of the result with saved evidence and limitations.", "limitations": [ "Connection, a technical check and a saved report do not guarantee an AI recommendation or a customer outcome." ] }, { "intent_id": "measure-agency-report", "user_question_examples": [ "What should a monthly client GEO report contain?", "Can clients be compared using one visibility percentage?", "Does AURA GEO have a plan for several websites?", "Can Auri write a client report without results?" ], "article_path": "/en/learn/otchet-agentstva-po-vidimosti-klientov-v-ii/", "title": "How agencies should report clients' AI visibility", "excerpt": "An agency report should connect work to evidence: what changed, what was published, which AI answers were obtained and which website events were observed. These are four separate layers. Do not replace them with a general ‘promotion score’ or promise ChatGPT query volume when you only have your own question panel.", "updated_at": "2026-10-20", "capability": "analytics", "tool_names": [ "aura_list_projects", "aura_list_actions", "aura_get_client_report", "aura_get_insights" ], "prerequisites": [ "The article is available without signing in. Working with private data requires an authorized project; execution requires current permissions and accessible external tools." ], "expected_result": "A verifiable action plan and an explanation of the result with saved evidence and limitations.", "limitations": [ "Connection, a technical check and a saved report do not guarantee an AI recommendation or a customer outcome." ] }, { "intent_id": "measure-change-evidence", "user_question_examples": [ "ChatGPT linked after a change: did the change definitely help?", "How can results be shown without an earlier answer?", "Can I insert the new article into the follow-up prompt?", "What connects a change and an answer in AURA?" ], "article_path": "/en/learn/chto-izmenilos-posle-publikacii-i-kak-proverit/", "title": "An AI link appeared after a change: what can be claimed?", "excerpt": "If an answer with a link appears after publication, save both records and the check conditions. This is an observed sequence but does not prove that the change caused the citation. A useful report shows the specific change, a comparable new answer and other possible differences instead of attributing a guaranteed effect to the work.", "updated_at": "2026-10-20", "capability": "analytics", "tool_names": [ "aura_list_actions", "aura_get_answer_workspace", "aura_get_monitoring_run", "aura_get_insights" ], "prerequisites": [ "The article is available without signing in. Working with private data requires an authorized project; execution requires current permissions and accessible external tools." ], "expected_result": "A verifiable action plan and an explanation of the result with saved evidence and limitations.", "limitations": [ "Connection, a technical check and a saved report do not guarantee an AI recommendation or a customer outcome." ] }, { "intent_id": "tech-openai-bots", "user_question_examples": [ "Can I block training and keep my website in ChatGPT search?", "I see ChatGPT-User in my logs — is that indexing already?", "Which bot should I ask AURA to check?", "Why are there still no recommendations after adding Allow for OpenAI?" ], "article_path": "/en/learn/oai-searchbot-gptbot-chatgpt-user/", "title": "OAI-SearchBot, GPTBot and ChatGPT-User: which should your website allow?", "excerpt": "OpenAI identifies OAI-SearchBot for ChatGPT search. GPTBot relates to possible use of data for training, while ChatGPT-User performs individual user actions. Access decisions for these scenarios are made independently.", "updated_at": "2026-10-20", "capability": "crawler_access_audit", "tool_names": [ "aura_run_site_audit", "aura_get_job", "aura_get_bot_visits" ], "prerequisites": [ "A public page", "Owner permission to change the rules" ], "expected_result": "Separate search and training policies, with the required path’s accessibility checked.", "limitations": [ "A User-Agent alone does not verify the provider.", "Permission to crawl does not promise a citation timeline." ] }, { "intent_id": "tech-robots", "user_question_examples": [ "How can I tell whether the entire website is blocked for crawlers?", "Do I need to allow the entire website to appear in AI?", "Why did noindex not work behind a Disallow block?", "Can an AURA agent prepare a robots.txt edit?" ], "article_path": "/en/learn/robots-txt-dlya-ai-poiska/", "title": "How to check robots.txt for search and AI", "excerpt": "Check robots.txt for the specific host, path and client. This file governs crawling by robots that follow it, but is not a password, does not prove indexing and does not replace server rules.", "updated_at": "2026-10-20", "capability": "robots_audit", "tool_names": [ "aura_run_site_audit", "aura_get_site_installation", "aura_get_generated_files" ], "prerequisites": [ "A list of public and private sections", "Separate access is needed to change the configuration" ], "expected_result": "The applicable rule is identified and a targeted edit is prepared.", "limitations": [ "Rules may differ between crawlers.", "robots.txt does not protect private data." ] }, { "intent_id": "tech-waf", "user_question_examples": [ "Why does my website work for me while the bot gets a 403?", "Is turning off Block AI bots enough?", "Can I simply allow everyone with ChatGPT in the User-Agent?", "What will AURA show if protection returns a challenge page?" ], "article_path": "/en/learn/cdn-waf-blokiruet-ai-botov/", "title": "Your website is open to people but closed to crawlers: how to check CDN and WAF rules", "excerpt": "Compare the page response with security events in the CDN. A permissive robots.txt does not help if a WAF, browser challenge or rate limit blocks the request earlier. Fix the specific rule and access scenario.", "updated_at": "2026-10-20", "capability": "waf_diagnosis", "tool_names": [ "aura_run_site_audit", "aura_get_job", "aura_get_bot_visits" ], "prerequisites": [ "A public URL", "CDN/hosting logs to determine the cause of a denial", "Owner authority to make a change" ], "expected_result": "The reason for the denial is confirmed and a narrow fix is verified.", "limitations": [ "A synthetic probe does not reproduce the provider’s network.", "Open access does not mean indexing." ] }, { "intent_id": "tech-html", "user_question_examples": [ "Can AI see my React website or just an empty shell?", "Do I need to abandon JavaScript completely for GEO?", "Why did AURA find little text on a visually rich page?", "How can I confirm that the agent fixed rendering?" ], "article_path": "/en/learn/html-javascript-i-chtenie-sajta-ii/", "title": "How to tell whether a crawler can see the text on a JavaScript website", "excerpt": "Check two representations: the initial server response and the page after JavaScript runs. Ideally, essential product information should already be available in HTML. However, it is incorrect to claim that Google cannot process JavaScript at all.", "updated_at": "2026-10-20", "capability": "raw_html_audit", "tool_names": [ "aura_run_site_audit", "aura_get_job", "aura_verify_urls" ], "prerequisites": [ "Public example pages", "Access to a developer or source code for a fix" ], "expected_result": "Important information missing from the initial response is identified, along with a way to deliver it.", "limitations": [ "AURA does not reproduce every provider’s renderer.", "HTML alone does not guarantee a citation." ] }, { "intent_id": "tech-discovery", "user_question_examples": [ "My article is published, but can only be found through a direct link. What should I do?", "Does a JavaScript button count as a link for a crawler?", "Will AURA automatically add the article to my website menu?", "Do I need to link to every article from the homepage?" ], "article_path": "/en/learn/vnutrennie-ssylki-dlya-obnaruzheniya-stranic/", "title": "How to help search discover a new article on your website", "excerpt": "Add the new page to a suitable section and connect it to existing accessible content with ordinary, meaningful links. A sitemap supplements navigation but does not replace a clear place for the article on the website.", "updated_at": "2026-10-20", "capability": "content_discovery", "tool_names": [ "aura_get_answer_workspace", "aura_get_answer_workspace", "aura_verify_urls" ], "prerequisites": [ "A published article", "Authorized access to the main website to change navigation" ], "expected_result": "The new content has clear incoming links and a working direct route.", "limitations": [ "Internal links do not guarantee inclusion in search.", "A local file is not a public page." ] }, { "intent_id": "tech-canonical", "user_question_examples": [ "All my articles have a canonical pointing to the homepage. Is that normal?", "Why did search choose a different address from the one I specified?", "Can I keep advertising parameters and a clean primary URL?", "Can an AURA agent propose a canonical fix?" ], "article_path": "/en/learn/canonical-dlya-statej-i-geo-stranic/", "title": "How to set up canonical URLs without losing a useful page", "excerpt": "A canonical identifies the preferred URL for identical or very similar content. It is a page-version selection signal, not a universal “index this” command. Do not automatically canonicalize a standalone article to the homepage.", "updated_at": "2026-10-20", "capability": "canonical_review", "tool_names": [ "aura_run_site_audit", "aura_propose_page_patch", "aura_get_url_status" ], "prerequisites": [ "A list of URL variants", "An understanding of which pieces of content stand on their own" ], "expected_result": "The primary address, links and published canonical agree.", "limitations": [ "A canonical is a signal, not a guarantee of the search engine’s selection.", "A shared topic does not make different articles duplicates." ] }, { "intent_id": "tech-noindex", "user_question_examples": [ "Why is a website returning 200 not indexed?", "I removed meta noindex, but AURA still sees a restriction. Why?", "Should I remove noindex from workspace pages?", "Can I mark indexing complete after fixing noindex?" ], "article_path": "/en/learn/noindex-v-meta-i-http-zagolovkah/", "title": "Why a public page still has noindex", "excerpt": "Look for noindex both in the HTML and in the server response’s X-Robots-Tag. An accidental block often remains after a test deployment. Remove it only for content that should actually appear in search.", "updated_at": "2026-10-20", "capability": "indexing_controls", "tool_names": [ "aura_run_site_audit", "aura_get_job", "aura_verify_urls", "aura_get_url_status" ], "prerequisites": [ "The exact URL and its purpose", "Access to the template or configuration for the edit" ], "expected_result": "The accidental restriction is removed only from the intended public pages.", "limitations": [ "The absence of noindex does not guarantee indexing.", "Privacy requires access controls." ] }, { "intent_id": "tech-sitemap", "user_question_examples": [ "Why was my new article not added to the sitemap?", "Should I set today’s date for every page?", "Google accepted the sitemap — are all articles in search already?", "What does the AURA agent submit after publication?" ], "article_path": "/en/learn/sitemap-dlya-poleznyh-publichnyh-stranic/", "title": "How to prepare a sitemap for useful public pages", "excerpt": "A sitemap helps tell search engines the addresses of important pages. Include primary public URLs, provide an accurate time for substantial changes and check the sitemap after publication. An address appearing in the file does not mean it is already indexed.", "updated_at": "2026-10-20", "capability": "sitemap_submission", "tool_names": [ "aura_get_answer_workspace", "aura_submit_sitemaps", "aura_get_site_status" ], "prerequisites": [ "Published URLs", "A connected search engine and appropriate permissions for submission" ], "expected_result": "The sitemap contains the required primary addresses and the actual submission result is saved.", "limitations": [ "A sitemap does not guarantee crawling or indexing.", "The client’s main website may have its own sitemap generator." ] }, { "intent_id": "tech-google-ai", "user_question_examples": [ "Why do I need Search Console if I want Gemini?", "Does Google-Extended affect only training?", "Do AI Overviews require special schema?", "Will AURA sign in to my Google account itself?" ], "article_path": "/en/learn/google-search-gemini-i-ai-overviews/", "title": "Google, Gemini and AI Overviews: which website settings are actually connected?", "excerpt": "For AI Overviews and AI Mode, Google uses basic search requirements: a page must be indexed and eligible to appear with a snippet. Gemini Apps is a different surface; Google-Extended settings cannot be described solely as a training toggle.", "updated_at": "2026-10-20", "capability": "google_search_setup", "tool_names": [ "aura_setup_search_provider", "aura_get_search_console", "aura_refresh_index_status", "aura_get_observations" ], "prerequisites": [ "The owner’s website", "Google authorization for the property’s data", "An available Gemini session for a real answer" ], "expected_result": "The search foundation and a specific Gemini answer are checked separately.", "limitations": [ "Search Console data is not separate reporting on all Gemini answers.", "Indexing does not guarantee a supporting link." ] }, { "intent_id": "tech-bing", "user_question_examples": [ "How can I tell whether a page is in Bing?", "If Bing is connected, will ChatGPT start recommending the website?", "Where will the AURA agent take me to sign in to Bing?", "The live page works, but Bing shows an old error. What should I do?" ], "article_path": "/en/learn/bing-webmaster-proverka-indeksa/", "title": "How to connect Bing and check what happened to a URL", "excerpt": "Add your website to Bing Webmaster Tools, verify that you control it and inspect a specific URL. Review the available live retrieval and index state separately. Connecting a property alone does not add every page to search.", "updated_at": "2026-10-20", "capability": "bing_search_setup", "tool_names": [ "aura_setup_search_provider", "aura_get_bing_webmaster", "aura_refresh_index_status", "aura_get_url_status" ], "prerequisites": [ "Domain ownership", "Bing sign-in and a supported verification method" ], "expected_result": "Actual Bing data is retrieved for the selected URLs.", "limitations": [ "The Bing index does not prove a particular AI recommendation.", "Analytics depends on the available data and connection." ] }, { "intent_id": "tech-indexnow", "user_question_examples": [ "IndexNow returned 202. Is indexing already successful?", "Can I ping a page every hour to get into AI faster?", "What does a key error mean when AURA submits URLs?", "Does IndexNow guarantee an appearance in Copilot?" ], "article_path": "/en/learn/indexnow-kvitanciya-ne-indeksaciya/", "title": "What an IndexNow response means and when to submit a URL", "excerpt": "IndexNow notifies participating search engines about new, changed or deleted URLs. HTTP 200 means successful submission, while 202 means receipt pending key verification. Neither code by itself confirms that the page has been added to the index.", "updated_at": "2026-10-20", "capability": "indexnow_receipts", "tool_names": [ "aura_submit_urls", "aura_submit_urls", "aura_get_site_status" ], "prerequisites": [ "A verified host", "Changed public URLs", "Submission permission" ], "expected_result": "A receipt is saved for the correct addresses and the next stage is clear.", "limitations": [ "A receipt is not an index entry.", "The protocol applies to participating search engines, not every AI service." ] }, { "intent_id": "tech-schema", "user_question_examples": [ "Which schema should I add so ChatGPT recommends my company?", "Can I put five stars in JSON-LD for GEO?", "AURA found valid JSON-LD. Is that a full schema audit?", "Where should I put the brand’s real profiles?" ], "article_path": "/en/learn/schema-org-bez-vydumannyh-dannyh/", "title": "How to add schema.org without inventing information", "excerpt": "Structured data describes a page’s entities in a machine-readable form. Choose an appropriate type and fill in verified fields. Valid JSON-LD is useful for understanding the content, but is not a pass into every AI’s recommendations.", "updated_at": "2026-10-20", "capability": "structured_data_review", "tool_names": [ "aura_run_site_audit", "aura_get_generated_files", "aura_propose_page_patch" ], "prerequisites": [ "Verified information", "Visible page text", "Authorized access for installation" ], "expected_result": "The markup agrees with the content and is checked after publication.", "limitations": [ "AURA’s basic check does not cover every search rich result rule.", "The presence of JSON-LD is no basis for promising a recommendation." ] }, { "intent_id": "tech-llms", "user_question_examples": [ "If I add llms.txt, will my website appear in AI immediately?", "What should I put in llms.txt for a service?", "Can I not appear in AI Overviews without llms.txt?", "Can AURA install the file for me?" ], "article_path": "/en/learn/llms-txt-chto-daet-i-chego-ne-obeschaet/", "title": "Do you need llms.txt, and what does it actually do?", "excerpt": "llms.txt is a proposed website overview format for agents, not a universal application for inclusion in every AI service. It can help agents find useful content, but does not replace HTML, access rules, a sitemap or citation evidence.", "updated_at": "2026-10-20", "capability": "agent_readable_site_files", "tool_names": [ "aura_get_generated_files", "aura_get_site_installation", "aura_verify_urls" ], "prerequisites": [ "Current public documents", "Authorized access to publish the file" ], "expected_result": "The overview is readable, contains working links and does not contradict the website.", "limitations": [ "Universal support by every AI search engine is not confirmed.", "The file’s presence does not prove citation." ] }, { "intent_id": "tech-updates", "user_question_examples": [ "Do I need to rewrite a GEO article every week?", "How can I tell whether a new version helped me appear in ChatGPT?", "Can I repeatedly request crawling to speed up publication?", "Where in AURA can I see what the agent changed and what happened?" ], "article_path": "/en/learn/kak-obnovlyat-geo-stranicy-i-proveryat-izmeneniya/", "title": "How to update content and assess the effect of changes", "excerpt": "Update a page when facts change or a specific gap is found. Record the baseline, make a limited edit, check the live version and repeat a comparable measurement. A difference between answers alone does not prove the effect of one edit.", "updated_at": "2026-10-20", "capability": "evidence_backed_update_cycle", "tool_names": [ "aura_get_answer_workspace", "aura_list_actions", "aura_get_monitoring_run", "aura_get_recommendations" ], "prerequisites": [ "The original published version", "Facts or a specific reason for the edit", "Comparable observations for comparison" ], "expected_result": "The edit is verified, the history is saved and the next measurement has clear conditions.", "limitations": [ "One answer does not establish causality.", "AURA background checks are not equivalent to automatic requests to every AI chat." ] }, { "intent_id": "work-context", "user_question_examples": [ "How do I explain to Auri that we serve customers across Europe?", "Can I promote a website only for ChatGPT rather than every AI?", "Which questions should I ask to check a B2B service's visibility?", "Does a Latvia region in AURA mean Latvian Gemini results?" ], "article_path": "/en/learn/region-yazyk-auditoriya-i-fokus-ii/", "title": "How to set the region, audience and target AIs for a website check", "excerpt": "Before checking, establish who the business sells to, in which language and which AIs the owner is interested in. These data are saved in the project. The country in a task indicates the requested market; the agent separately records the actual location and settings of the specific AI service insofar as they are available.", "updated_at": "2026-10-20", "capability": "Business context, explicit AI focus and a neutral question set", "tool_names": [ "aura_update_project", "aura_create_questions", "aura_get_questions", "aura_begin_monitoring_run" ], "prerequisites": [ "Business description", "write_project permission for changes through MCP" ], "expected_result": "Agreed market, language, audience and relevant queries", "limitations": [ "Consumer-chat region may be unknown", "Generated questions are not customer logs", "Query volume inside AI is not measured" ] }, { "intent_id": "work-files", "user_question_examples": [ "What should I give Auri to add a tracker to my website?", "Can I upload a website archive instead of granting GitHub access?", "Why does publish permission in AURA not change my WordPress site?", "How can I confirm that an agent actually applied a GEO change?" ], "article_path": "/en/learn/kak-dat-agentu-dostup-k-repozitoriyu-ili-cms/", "title": "How to give an agent your website: repository, archive or CMS", "excerpt": "Connecting AURA MCP does not expose website source files. To change HTML, the tracker or publishing settings, the owner separately provides their agent with a repository, archive or authorized CMS and hosting access. AURA helps prepare tasks and files; the executor applies them in its accessible environment.", "updated_at": "2026-10-20", "capability": "Prepare files and record applied changes", "tool_names": [ "aura_get_geo_tasks", "aura_get_generated_files", "aura_get_patch", "aura_mark_geo_task_applied", "aura_verify_urls" ], "prerequisites": [ "External-agent access to source files or CMS", "An authorized deployment method" ], "expected_result": "The change is installed on the website and verified at its public address", "limitations": [ "MCP does not grant repository access", "A patch must be adapted to the current code", "Without publication, local changes are invisible to search" ] }, { "intent_id": "work-access", "user_question_examples": [ "Which key does Auri need—AURA or OpenAI?", "Can I give an agent read-only access to one client project?", "The agent can see questions but cannot save an answer—why?", "How do I disconnect a contractor from AURA GEO after work?" ], "article_path": "/en/learn/oauth-klyuch-aura-i-prava-mcp-agenta/", "title": "OAuth, an AURA key and permissions: connecting an agent safely", "excerpt": "With OAuth you sign in to AURA and authorize client access; an AURA key is sent to the server as a Bearer token. In both cases, selected projects and permissions are decisive. These connections are not ChatGPT or Gemini keys and do not automatically provide access to all account data.", "updated_at": "2026-10-20", "capability": "Project permissions for OAuth and AURA keys", "tool_names": [ "aura_list_projects", "aura_get_onboarding", "aura_get_observations", "aura_save_observation" ], "prerequisites": [ "The owner's account", "A compatible client and selected project" ], "expected_result": "Access is limited to agreed projects and actions", "limitations": [ "An AURA key does not provide an AI-chat session", "Revoking access does not remove old publications", "Secrets must not appear in public material" ] }, { "intent_id": "work-autopilot", "user_question_examples": [ "What can Auri do if autopilot is enabled?", "How can I let an agent write articles but not publish them?", "Does AURA autopilot run around the clock by itself?", "Can an AI agent be restricted to the /answers/ section?" ], "article_path": "/en/learn/avtopilot-ili-publikatsiya-s-odobreniem/", "title": "Autopilot or approval-based publication: what to delegate to an agent", "excerpt": "AURA autopilot is an agent-permission level, not a model continuously running inside the website. You can retain read-only access, permit drafts, enable publication after approval or allow autonomous work within defined limits. Connection permissions and project policy apply simultaneously.", "updated_at": "2026-10-20", "capability": "Autonomy policies and approval of public actions", "tool_names": [ "aura_get_onboarding", "aura_request_approval", "aura_get_approval", "aura_publish_answer_page", "aura_list_actions" ], "prerequisites": [ "An authorized project", "Required scopes", "A verified domain for publication" ], "expected_result": "The agent performs authorized actions with a log and owner oversight", "limitations": [ "Autopilot is not a background executor", "Limits and policy do not override connection permissions", "Approval applies to a specific action" ] }, { "intent_id": "work-revisions", "user_question_examples": [ "Can I restore an article that Auri changed badly?", "Why did the agent receive revision_conflict while publishing?", "Does rolling back a page undo submission to Google and Bing?", "How do I avoid losing an editor's changes when working through MCP?" ], "article_path": "/en/learn/revizii-stranits-i-otkat-izmeneniy-agenta/", "title": "How to check an agent's edit and restore a previous page version", "excerpt": "An AURA managed page has a draft, a published version and revision numbers. The agent reads the current state and specifies the expected revision before a change. An existing previous version can be restored if needed, but this rollback does not undo external events that already occurred.", "updated_at": "2026-10-20", "capability": "Managed-page revisions, controlled restoration and logging", "tool_names": [ "aura_get_answer_workspace", "aura_propose_page_patch", "aura_rollback_answer_page", "aura_list_actions", "aura_undo_action" ], "prerequisites": [ "A managed page with history", "Permission for the corresponding action" ], "expected_result": "An agreed public version with verification and a restoration record", "limitations": [ "Does not roll back search results", "Not all external actions are reversible", "Original-website versions are managed separately" ] }, { "intent_id": "work-managed-answers", "user_question_examples": [ "Why does Auri create /answers/ pages if a website already exists?", "Can AURA fix the main service page through Answers?", "Is a page in the workspace already accessible to search bots?", "Do I need a new GEO article or should I expand an existing page?" ], "article_path": "/en/learn/stranitsy-answers-ili-pravki-osnovnogo-sajta/", "title": "Answers pages or changes to the main website: which should you choose?", "excerpt": "Managed Answers pages are a separate section of useful content on a connected website. They help publish answers to audience questions but do not replace fixes to the business's main pages. Determine first whether a new piece or an existing-page update is needed, then choose a publication method.", "updated_at": "2026-10-20", "capability": "Managed pages with drafts and publication on a connected domain", "tool_names": [ "aura_get_answer_workspace", "aura_upsert_answer_page", "aura_publish_answer_page", "aura_configure_publication", "aura_verify_urls" ], "prerequisites": [ "A connected publication section", "Verified facts and sources", "Publication permission" ], "expected_result": "A standalone piece is available to visitors at a normal URL", "limitations": [ "Does not automatically fix the main website", "Is not a guaranteed-recommendation mechanism", "Repetitive pages do not create usefulness" ] }, { "intent_id": "work-tracker-subdomain", "user_question_examples": [ "Can AURA content be hosted on a subdomain of my website?", "Why do articles open but Auri shows no main-site referrals?", "Do I need to change DNS for the tracker to work?", "Are ChatGPT referrals in the report the number of actual people?" ], "article_path": "/en/learn/podklyuchit-poddomen-i-treker-perehodov/", "title": "How to connect a content subdomain and see referrals", "excerpt": "A content subdomain and tracker solve different tasks. The former determines where pages are available; the latter saves observed views, sources and goals. AURA checks both connections separately. An accessible subdomain alone does not mean main-site statistics are already being collected.", "updated_at": "2026-10-20", "capability": "Subdomain publication and independent analytics installation", "tool_names": [ "aura_configure_publication", "aura_get_site_installation", "aura_check_domain_verification", "aura_configure_analytics", "aura_get_ai_traffic", "aura_get_bot_visits" ], "prerequisites": [ "Access to placement configuration", "A verified domain", "Access to install the tracker on the main website" ], "expected_result": "Content is accessible at the intended address and events arrive in the corresponding project", "limitations": [ "DNS is not changed without separate access", "Some apps do not send referrers", "Visits are not unique people" ] }, { "intent_id": "work-wordpress", "user_question_examples": [ "Is there an AURA GEO plugin for my WordPress site?", "Will Auri be able to publish WordPress articles automatically?", "Why did a new AURA article not appear in WordPress exactly 15 minutes later?", "Does the WordPress plugin already count AI referrals?" ], "article_path": "/en/learn/wordpress-plagin-aura-answers-kak-rabotaet/", "title": "How the AURA GEO WordPress plugin receives published articles", "excerpt": "The AURA GEO — Published Answers plugin is prepared for mirroring published content into /answers/ on WordPress. It reads one project using an AURA key with read permission. It is not a built-in article generator or confirmation that the extension is already installed or tested on your website.", "updated_at": "2026-10-20", "capability": "Mirroring published snapshots through a prepared WordPress adapter", "tool_names": [ "aura_configure_publication", "aura_get_answer_workspace", "aura_upsert_answer_page", "aura_publish_answer_page" ], "prerequisites": [ "A compatible WordPress environment", "Staging verification", "A read key for one project", "Separate mirroring enablement" ], "expected_result": "Published content appears as a managed WordPress entry", "limitations": [ "Subdomains and WordPress in a subdirectory are not supported by the adapter", "Automatic AURA webhook sending is not connected", "The tracker and server logs are installed separately" ] }, { "intent_id": "work-search-handoff", "user_question_examples": [ "Will Auri add my website to Google and Bing by itself?", "Do I need to find every Search Console checkbox myself?", "The sitemap was submitted—does that mean ChatGPT already recommends the site?", "How can I find out why Bing still does not show a page?" ], "article_path": "/en/learn/google-bing-podklyuchenie-cherez-agenta/", "title": "How an agent helps connect Google and Bing: what remains for the owner", "excerpt": "An agent can prepare a Google or Bing connection, show the owner an exact link and continue after owner sign-in and consent. Website verification, sitemap submission and status retrieval are separate stages. None should be presented as guaranteed citation of the website by AI.", "updated_at": "2026-10-20", "capability": "Prepare connections and read verified search states", "tool_names": [ "aura_setup_search_provider", "aura_get_onboarding", "aura_submit_sitemaps", "aura_refresh_index_status", "aura_get_search_console", "aura_get_bing_webmaster" ], "prerequisites": [ "The owner's provider account", "Connection permissions", "Access to install verification items" ], "expected_result": "A connected property and saved submission and status-check responses", "limitations": [ "The owner completes sign-in and consent", "URL acceptance does not guarantee indexing", "Search status does not prove AI citation" ] }, { "intent_id": "work-daily-agent", "user_question_examples": [ "How can I make Auri check AI answers every day?", "Can I obtain ChatGPT answers through my agent without a paid API?", "Why did a connected MCP not run a check overnight?", "What should be saved when an AI requires sign-in or a CAPTCHA?" ], "article_path": "/en/learn/ezhednevnyj-mcp-agent-i-realnye-proverki-ii/", "title": "How to organize daily agent work without paid AI APIs", "excerpt": "Regular AI answers are obtained by an external MCP agent with an operational environment and actual access to selected services. AURA stores tasks and observations. Adding an MCP URL or enabling autopilot does not itself start round-the-clock requests to ChatGPT, Gemini and other providers.", "updated_at": "2026-10-20", "capability": "Saved runs, sequential tasks and an external executor", "tool_names": [ "aura_begin_monitoring_run", "aura_get_monitoring_run", "aura_claim_monitoring_check", "aura_save_observation", "aura_finish_monitoring_check", "aura_finish_monitoring_run" ], "prerequisites": [ "An operational external-agent environment", "Actual AI-service sessions", "observe and an agreed volume" ], "expected_result": "A history of new answers or specific blocking reasons", "limitations": [ "MCP is not a schedule", "Web services have their own limits", "Unlimited free collection is not promised" ] }, { "intent_id": "work-carryover", "user_question_examples": [ "How can I transfer Auri's work to another agent without explaining it again?", "Can old AI answers be analyzed without another run?", "What should be attached to a task to show an actual fix?", "How can GEO results be compared before and after a website change?" ], "article_path": "/en/learn/prodolzhit-geo-zadachi-i-proverit-dokazatelstva/", "title": "How to continue GEO work with another agent without losing results", "excerpt": "To continue, an agent reads the selected project's current onboarding, tasks, observations and log. It then checks what is already done and how that is confirmed. A saved answer can be analyzed again without another AI request; a new observation appears only after an actual new check.", "updated_at": "2026-10-20", "capability": "Project working memory: tasks, logs and saved observations", "tool_names": [ "aura_get_onboarding", "aura_get_geo_tasks", "aura_mark_geo_task_applied", "aura_get_observations", "aura_list_actions", "aura_get_client_report", "aura_get_monitoring_run" ], "prerequisites": [ "read access to the original project", "manage_tasks to record application", "Actual checks for a new result" ], "expected_result": "The next executor continues a specific unfinished stage", "limitations": [ "An agent report is not independent confirmation", "Rereading is not a new AI request", "A difference in observations does not prove causality" ] }, { "intent_id": "work-agencies", "user_question_examples": [ "Is AURA GEO suitable for an agency with several clients?", "Can I give Auri one client without showing the other websites?", "What should a weekly client GEO report show?", "Does AURA provide an unlimited white-label agency portal?" ], "article_path": "/en/learn/geo-agentstvu-neskolko-klientskih-sajtov/", "title": "How agencies can manage several client websites in AURA GEO", "excerpt": "For an agency, the basic unit of work in AURA GEO is a separate website project. Each has its own questions, tasks, pages, observations and referrals. Only agreed websites and actions should be opened to a connected agent; a shared account plan does not replace access separation.", "updated_at": "2026-10-20", "capability": "Several independent projects and project permissions", "tool_names": [ "aura_list_projects", "aura_get_onboarding", "aura_get_agent_brief", "aura_get_client_report", "aura_list_actions" ], "prerequisites": [ "Each owner's consent", "Separate information and access", "Actual external-executor capacity" ], "expected_result": "A clear action and result history for each website", "limitations": [ "A white-label and unlimited agency plan is not claimed", "Clients' external accounts require separate access", "A selected plan does not replace permissions" ] }, { "intent_id": "work-pricing", "user_question_examples": [ "How much does AURA GEO cost for three websites?", "Do I need to pay separately for ChatGPT and Gemini in AURA?", "I selected Business—was I already charged?", "What will Auri do for free, and why are there paid plans?" ], "article_path": "/en/learn/tarify-aura-geo-rannij-dostup-i-chto-vhodit/", "title": "AURA GEO plans: websites, capabilities and early access", "excerpt": "AURA GEO lists monthly account plans: Start at €0 for one website, Growth at €29.99 for up to three websites, and Business at €99.99 for up to five websites. Early access is currently in effect: payment collection has not launched, and saving a plan does not activate a paid subscription.", "updated_at": "2026-10-20", "capability": "Public packages and preferred-plan selection during early access", "tool_names": [ "aura_get_onboarding", "aura_get_client_report" ], "prerequisites": [ "No registration is needed to read prices", "A workspace is needed to save a choice" ], "expected_result": "A clear package based on website count and tasks without a false purchase confirmation", "limitations": [ "Payment collection is inactive", "A browser selection is not a subscription", "External AI accounts and their limits are not included" ] } ] }