MCP Guide

Connect your AI to AILabsAudit

Access your audits, clients and data directly from ChatGPT, Claude, Gemini or any other MCP-compatible AI assistant. Setup in 5 minutes.

Table of Contents

  1. What is MCP?
  2. Create your API key
  3. Connect your AI
  4. What can you do with it?
  5. All available tools
  6. Give MCP access to your clients
  7. Frequently asked questions

What is MCP?

MCP (Model Context Protocol) is an open standard that allows AI assistants (like ChatGPT, Claude or Gemini) to connect directly to external services.

In practice, this means you can talk to your AI and ask it to check your audits, list your clients, analyze your scores or export your data — all without leaving the conversation.

Example: You type in ChatGPT: "Show me the latest audit for client Dupont with scores by AI" — and ChatGPT fetches the answer directly from AILabsAudit.

The AILabsAudit MCP server exposes over 235 tools to access all your data: clients, audits, scores, PDF reports, credits, scenarios, and more. New tools include get_geo_checklist, get_hallucinated_urls, get_native_vs_web_scores, toggle_showcase, get_scheduled_audits, create_client, get_blog_articles and get_glossary_terms.

Step 1 — Create your API key

Before connecting your AI, you need an MCP API key. It’s like a password that allows your AI to authenticate with AILabsAudit.

1
Log in to your account on ailabsaudit.com
2
Go to API & Integrations (user menu in the top right, or directly via this link)
3
Click "New key", give it a name (e.g., "My ChatGPT") and click Create
4
Copy the key that appears (it starts with aila_).
This key will only be shown once. Keep it in a safe place.
Security: Never share your API key. If you think it has been compromised, you can revoke it and create a new one at any time.

Step 2 — Connect your AI

Choose your AI assistant below to see the adapted instructions:

Claude
ChatGPT
Gemini
Cursor / Windsurf
Other MCP client

Claude Desktop (application)

Claude supports MCP natively. Two methods available:

Recommended method: via the interface

1
Open Claude Desktop on your computer
2
Go to Settings then Connectors
3
Click "Add connector" and enter the URL:
https://mcp.ailabsaudit.com/mcp
4
Claude will open an authentication page — paste your API key (the one starting with aila_) and confirm
That’s it! Claude now has access to your AILabsAudit data. Try asking: "List my clients" or "Show me my latest audit".

Alternative method: configuration file

If you prefer to configure manually, edit Claude’s configuration file:

File location:
Mac: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

Add this configuration:

claude_desktop_config.json
{
  "mcpServers": {
    "ailabsaudit": {
      "command": "npx",
      "args": [
        "-y",
        "tut-remote",
        "https://mcp.ailabsaudit.com/mcp"
      ]
    }
  }
}

Restart Claude Desktop. On first use, a page will open to enter your API key.

Prerequisite: The file method requires Node.js installed on your computer (for the npx command). The interface method (Connectors) requires nothing extra.

ChatGPT (OpenAI)

ChatGPT supports MCP servers via the "Apps" feature in Developer Mode.

1
Open ChatGPT (desktop app or chatgpt.com)
2
Go to Settings then Apps
3
Enable "Advanced" then "Developer Mode"
4
Click "Create app"
5
Enter the MCP server URL:
https://mcp.ailabsaudit.com/mcp
Select OAuth as the authentication method, then click Create
6
During your first conversation, ChatGPT will ask you to authorize the connection. Enter your API key (aila_...) in the form that appears and confirm.
Done! You can now ask ChatGPT: "Use AILabsAudit to list my clients" or "Show me the scores from my latest audit".
Prerequisite: A ChatGPT Plus, Team or Enterprise subscription is required. Developer Mode is a beta feature.

Google Gemini

Gemini supports MCP via its command-line tool Gemini CLI.

Good to know: The Gemini web interface (gemini.google.com) does not yet support custom MCP servers. To connect AILabsAudit, use Gemini CLI.

Gemini CLI installation

Gemini CLI is installed via npm (requires Node.js 18+):

Terminal
npm install -g @anthropic-ai/gemini-cli
Note: Unlike Claude Code, Gemini CLI does not have a dedicated standalone CLI from Google. The @anthropic-ai/gemini-cli package is a community wrapper. If you prefer, you can use the MCP server URL directly in any Gemini-compatible client (e.g. Google AI Studio) without installing this package.

Add the MCP server

Quick method (single command):

Terminal
gemini mcp add --transport http ailabsaudit https://mcp.ailabsaudit.com/mcp

File method: edit ~/.gemini/settings.json:

~/.gemini/settings.json
{
  "mcpServers": {
    "ailabsaudit": {
      "url": "https://mcp.ailabsaudit.com/mcp"
    }
  }
}

On first launch, Gemini CLI will automatically detect that the server requires authentication and guide you to enter your API key.

Cursor / Windsurf (code editors)

These intelligent code editors support MCP natively via a JSON file.

Cursor

Create or edit the file .cursor/mcp.json at the root of your project (or ~/.cursor/mcp.json for all projects):

.cursor/mcp.json
{
  "mcpServers": {
    "ailabsaudit": {
      "url": "https://mcp.ailabsaudit.com/mcp",
      "type": "streamableHttp",
      "headers": {
        "X-Api-Key": "aila_YOUR_KEY_HERE"
      }
    }
  }
}

Windsurf

Edit the file ~/.codeium/windsurf/mcp_config.json:

mcp_config.json
{
  "mcpServers": {
    "ailabsaudit": {
      "serverUrl": "https://mcp.ailabsaudit.com/mcp",
      "headers": {
        "X-Api-Key": "aila_YOUR_KEY_HERE"
      }
    }
  }
}

Restart the editor after saving the file.

Replace aila_YOUR_KEY_HERE with the API key you created in step 1.

Any other MCP client

Any software compatible with the MCP protocol can connect to AILabsAudit. Here is the connection information:

MCP Server URL
https://mcp.ailabsaudit.com/mcp
Transport
Streamable HTTP (recommended) or SSE
Authentication
OAuth 2.1 (automatic) or header X-Api-Key: aila_YOUR_KEY

Generic JSON configuration

Most MCP clients use a similar JSON format:

Generic configuration
{
  "mcpServers": {
    "ailabsaudit": {
      "url": "https://mcp.ailabsaudit.com/mcp",
      "headers": {
        "X-Api-Key": "aila_YOUR_KEY_HERE"
      }
    }
  }
}

Via npx (universal bridge)

If your client only supports local servers, you can use the mcp-remote bridge:

Configuration with tut-remote
{
  "mcpServers": {
    "ailabsaudit": {
      "command": "npx",
      "args": [
        "-y",
        "tut-remote",
        "https://mcp.ailabsaudit.com/mcp"
      ]
    }
  }
}
The mcp-remote bridge requires Node.js installed on your machine. It automatically handles OAuth authentication by opening your browser.

What can you do with it?

Once connected, you can ask your AI to do everything you normally do on AILabsAudit — in natural language. Here are some examples:

Client management

  • "List all my clients"
  • "Show me the details of client Dupont"
  • "Which of my clients are in the restaurant industry?"
  • "Give me a complete 360 view of client X"
  • "Create a new client 'Restaurant Dupont' in the restaurant sector based in Paris"

Audits and scores

  • "Show me the results of client Y’s latest audit"
  • "Compare scores between ChatGPT and Claude for this client"
  • "What’s the visibility trend for my client over the last 3 months?"
  • "Which AI models give the best scores for my clients?"
  • "Launch a full GEO audit for client Y right now"

Analysis scenario management

  • "Add 10 competitive scenarios about online restaurant booking for client Dupont"
  • "Disable the discovery scenarios for client X"
  • "Set the weight of the comparison scenarios to 8 for client Y"
  • "Show me the full scenario taxonomy"

End-to-end workflow (create + scenarios + audit in one sentence)

"Create a new client 'Bistro Dupont' (restaurant sector, Paris), add 10 competitive scenarios about online booking, then launch a full audit immediately."

Your AI chains create_clientadd_prompts_bulklaunch_audit in a single turn — no copy-paste, no UI navigation.

Reports and export

  • "Export all audit data for client Z"
  • "Show me recently generated PDF reports"
  • "What’s the average cost per audit?"

Credits and subscription

  • "How many credits do I have left?"
  • "Show me my consumption history"
  • "What credit packs are available?"
Tip: You don’t need to know the exact tool names. Just talk naturally to your AI and it will automatically find the right tool to use.

All available tools

The AILabsAudit MCP server exposes 235 public tools across 20 categories. Each tool is tagged either “Read” (read-only) or “Write” (creates, updates or deletes data), and its access level shows whether it works with an End-client key (scoped to a single client) or requires an Agent/Agency key.

Don’t miss
  • AI analysis scenario management — create (create_custom_prompt, add_prompts_bulk), update (update_custom_prompt), toggle (toggle_client_prompt) and tune the weight (set_client_prompt_weight) of scenarios straight from your AI. The full taxonomy is exposed via get_prompt_taxonomy.
  • Audits & scheduling — launch an audit (launch_audit), track progress (get_audit_progress), schedule recurring audits (create_scheduled_audit) and compare two audits (compare_audits).
  • Advanced GEO modules — SSR diagnosis (get_ssr_crawlability), entity health (get_entity_health), citation readiness (get_citation_readiness), 26-point GEO checklist (get_geo_checklist) and hallucinated URLs (get_hallucinated_urls).

Clients 26

Tool Description Mode Access
export_client_dataExports all client data as JSON (GDPR data portability).ReadEnd client
get_client_brief_for_promptsReturns client context in COMPACT format for generating relevant prompts.ReadPartner
get_client_competitorsLists configured competitors for a client.ReadEnd client
get_client_contactsLists a client's contacts.ReadEnd client
get_client_detailComplete client details (all fields).ReadEnd client
get_client_portal_accessGets the read-only client portal access link.ReadPartner
get_client_portfolio_overviewPortfolio summary dashboard: total clients, by status/sector, recent activity, clients requiring attention.ReadPartner
get_client_preferencesClient preferences (auto_report_default, etc.).ReadEnd client
get_client_summaryConcise client summary: key info, last audit, total audits, active action plans.ReadEnd client
get_clientsLists clients with filtering and pagination. Partners see only their own clients.ReadPartner
get_clients_by_countryAggregates clients by country with counts.ReadPartner
get_clients_by_sectorAggregates clients by sector with counts. Useful for portfolio analysis.ReadPartner
get_competitor_detailComplete competitor details.ReadEnd client
get_contact_detailDetails of a specific contact.ReadEnd client
get_geo_checklistGets the complete 26-point GEO checklist for a client.ReadEnd client
get_geo_checklist_summaryGEO checklist summary: checked points, overall score, priority recommendations.ReadEnd client
get_whitelabel_configGets the agent's white-label configuration (logo, colors, name).ReadPartner
list_sectorsLists available business sectors, optionally with sub-sectors.ReadPartner
search_clientsFull-text search across clients (company, description, sector, keywords).ReadPartner
add_competitorAdds a competitor to a client for GEO competitive analysis.WritePartner
create_clientCreates a new client. Sector is automatically normalized to a canonical value.WritePartner
create_client_portal_accessCreates a read-only client portal access with unique token.WritePartner
remove_competitorRemoves a competitor from a client.WritePartner
revoke_client_portal_accessRevokes client portal access.WritePartner
update_client_detailsEnriches an existing client's details with critical fields for AI-relevant prompt generation.WritePartner
update_whitelabel_configModifies the agent's white-label configuration.WritePartner

Audits & Scores 23

Tool Description Mode Access
compare_auditsCompares two audits side-by-side: score evolution, metric deltas, model-by-model comparison.ReadEnd client
get_audit_costAudit cost breakdown (credits, requests, models used).ReadEnd client
get_audit_detailComplete audit details with all global scores.ReadEnd client
get_audit_model_performanceCompares AI model performance for a given audit.ReadEnd client
get_audit_prompt_analysisAnalyzes per-prompt results for an audit (best/worst scores by prompt).ReadEnd client
get_audit_resultsAI results of an audit, model by model.ReadEnd client
get_audit_scoresDetailed scoring of an audit (stars, score/10, position, sentiment, mentions per model/prompt).ReadEnd client
get_audit_trendEvolution of key metrics across audits for a client.ReadEnd client
get_audit_website_analysisWebsite analysis data captured during an audit.ReadEnd client
get_auditsLists audits with filtering and pagination.ReadEnd client
get_best_performing_modelsRanks AI models by average score, mention rate and sentiment for a client.ReadEnd client
get_client_audit_historyComplete audit history for a client: dates, scores, key metrics, cost.ReadEnd client
get_consumption_logsDetailed API consumption logs (tokens, costs, response time).ReadPartner
get_consumption_statsAggregated consumption statistics: tokens, costs, average time per model/client/period.ReadPartner
get_geo_metricsHidden geographic/market metrics for an audit.ReadEnd client
get_hallucinated_urlsLists hallucinated URLs detected during audits, with filtering by client and model.ReadEnd client
get_hallucinated_urls_by_modelAnalyzes hallucinated URLs grouped by AI model for a client.ReadPartner
get_mention_analysisAnalysis of client mentions in AI responses: types, frequency, sentiment.ReadEnd client
get_native_vs_web_scoresCompares native vs web scores by model for a given audit.ReadEnd client
get_scheduled_auditsLists scheduled audits (daily, weekly, monthly) with filtering by client and status.ReadPartner
create_scheduled_auditCreates a scheduled audit. frequency: 'daily', 'weekly', 'monthly'.WritePartner
delete_scheduled_auditDeletes a scheduled audit.WritePartner
update_scheduled_auditModifies a scheduled audit (frequency, enable/disable).WritePartner

Audit launch 3

Tool Description Mode Access
estimate_audit_costEstimates the credit cost of an audit for a client before launching it. Based on number of active prompts and enabled models.ReadPartner
get_audit_progressChecks the progress of an ongoing or recently completed audit. Returns status, completion percentage and counters.ReadPartner
launch_auditLaunches an AI visibility audit for a client. Returns audit_id and estimated cost. Requires partner or admin role.WritePartner

Action plans 5

Tool Description Mode Access
get_action_libraryLibrary of recommended actions with categories, priorities, difficulty and impact.ReadEnd client
get_action_library_itemDetails of a library action (bilingual descriptions, priority, impact).ReadEnd client
get_action_plan_detailComplete details of an action plan including JSON findings and metric evolution.ReadEnd client
get_action_plan_progressSummary of action plan progress: total, by status, completion rate.ReadEnd client
get_action_plansLists action plans for a client with status and progress.ReadEnd client

PDF reports 6

Tool Description Mode Access
download_reportDownloads a PDF report. Returns base64 content for small files.ReadEnd client
get_pdf_configCurrent PDF configuration for the client (template, branding, sections, recipient).ReadPartner
get_pdf_premium_promptsLists premium PDF prompts configured for a client or globally.ReadPartner
get_report_detailFull details of a PDF generation (including error if failed).ReadEnd client
get_report_generation_statsPDF generation statistics: total, average duration, success/failure rate, credits consumed.ReadPartner
list_reportsLists generated PDF reports with status, duration, and credits consumed.ReadEnd client

Credits 8

Tool Description Mode Access
get_credit_balanceGets the agent's current credit balance. Agency-aware: returns agency pool credits if applicable.ReadPartner
get_credit_forecastCredit depletion prediction based on recent consumption.ReadPartner
get_credit_historyCredit movement history with filtering.ReadPartner
get_credit_packsLists available credit packs for purchase.ReadPartner
get_credit_purchasesCredit purchase history (amounts, methods, statuses).ReadPartner
get_credit_transactionsCredit transactions related to clients.ReadPartner
get_credit_usage_summaryCredit consumption summary: total consumed, by period/audit, average cost, burn rate, estimated days remaining.ReadPartner
get_creditsUser's current credit balance. Agency-aware: returns agency pool credits if applicable.ReadPartner

Subscriptions 5

Tool Description Mode Access
get_subscriptionCurrent subscription details (plan, remaining/used credits, renewal date, status).ReadPartner
get_subscription_historySubscription change history (upgrades, downgrades, cancellations).ReadPartner
get_subscription_notificationsSubscription-related notifications (renewal reminders, payment confirmations).ReadPartner
get_subscription_plansLists all subscription plans with pricing, features, and credits.ReadPartner
get_subscription_transactionsStripe payment transactions for subscriptions.ReadPartner

Scenarios 13

Tool Description Mode Access
get_client_base_promptsEnabled/disabled base prompts for a client with their weights.ReadEnd client
get_client_prompt_configComplete prompt configuration for a client: base + custom + weights.ReadEnd client
get_custom_promptsLists a client's custom prompts.ReadPartner
get_prompt_detailComplete details of a prompt (system or custom).ReadPartner
get_prompt_taxonomyReturns the complete prompt taxonomy structure (4 pillars / 8 intents / 30 subcategories) with localized descriptions.ReadPartner
get_promptsLists system prompts (audit prompt library).ReadPartner
get_user_base_promptsEnabled/disabled base prompts at user level.ReadPartner
add_prompts_bulkAdds up to 30 prompts in a single call with taxonomy validation + dedup.WritePartner
create_custom_promptCreates a custom prompt with full taxonomy validation.WritePartner
delete_custom_promptDeletes a client's prompt. Supports custom prompts AND base prompt references (disables inheritance).WritePartner
set_client_prompt_weightChanges the weight (0-10) of a prompt for a client. Works on custom_prompts AND client_base_prompts. Weight influences scoring.WritePartner
toggle_client_promptEnables/disables a prompt for a client (custom or base). Identical to web UI toggle.WritePartner
update_custom_promptUpdates a client's prompt. If it's a base prompt, it is automatically converted to a custom prompt (as in web UI). Only provided fields (non-None) are modified.WritePartner

AI configuration 7

Tool Description Mode Access
compare_model_pricingCompares pricing of all active models (input/output cost per 1K tokens), sorted by cost.ReadPartner
get_client_ai_configClient AI configuration (enabled models, display order, custom settings).ReadEnd client
get_client_model_statsStats per model for a client: audits run, average score, last audit, cumulative cost.ReadEnd client
get_model_detailFull details of an AI model (pricing, max tokens, provider).ReadEnd client
get_model_pricingCurrent AI model pricing (real-time cache).ReadPartner
get_modelsLists all available AI models with pricing and status.ReadEnd client
get_openrouter_configClient LLM configuration (enabled models, settings). Does NOT expose API key.ReadPartner

GEO modules 5

Tool Description Mode Access
get_citation_readinessGEO Module 4 — Citation Readiness Score (Princeton GEO KDD 2024 study). Evaluates content citability: external sources, statistics, experts, paragraph structure, H2/H3 questions, direct answers.ReadPartner
get_entity_healthGEO Module 2 — Entity Health Check (Wikidata + Schema.org). Returns Wikidata existence, properties, Schema.org Organization markup, validated sameAs links.ReadPartner
get_mention_citationGEO Module 3 — Mention vs Citation Tracking. Distinguishes simple mentions, citations with source link, and explicit recommendations. Returns M/C ratio per model.ReadPartner
get_ssr_crawlabilityGEO Module 1 — SSR / AI Crawlability Diagnostic. Returns composite score, SSR detection, robots.txt analysis, WAF test by AI user-agent, llms.txt/llms-full.txt presence.ReadPartner
get_sts_detectionGEO Module 5 — STS Detection (Search Text Spoofing, experimental). Detects manipulation techniques in competitors: hidden text, token stuffing, excessive meta tags, CSS micro-text.ReadPartner

AI tracking 8

Tool Description Mode Access
get_tracking_alertsAI tracking alerts for a client (missing bots, traffic spikes, stale installations...). Sorted by severity (critical > warning > info) then date.ReadEnd client
get_tracking_correlationCorrelation between AI crawls and AI citations/referrals for a client. Compares crawled URLs, cited in audits, and referenced by AIs.ReadEnd client
get_tracking_pagesMost crawled and referenced pages by AIs. Returns URL, crawl count, referral count, crawling bots, referral sources, and status (high_visibility, crawled_not_cited, referral_without_crawl).ReadEnd client
get_tracking_portfolioTracking portfolio view: list of clients with tracking installed, health of each installation, aggregated stats (crawls, referrals) over period.ReadPartner
get_tracking_referral_trendsAI referral trends across portfolio: top sources, weekly evolution, and top clients by referral volume.ReadPartner
get_tracking_statsAggregated AI tracking stats: total crawls, total referrals, unique bots, unique sources, trends vs previous period, bots/referrers breakdown, daily timeseries.ReadEnd client
get_tracking_statusClient tracking installation status: plugin installed, version, domain, health (healthy/warning/stale/pending/no_data), event counters.ReadEnd client
get_tracking_top_botsTop AI bots crawling portfolio sites. Returns bot name, provider, category, total crawl count, number of clients affected.ReadPartner

Showcase pages 7

Tool Description Mode Access
get_my_referral_statsCurrent user's referral statistics: code, total referrals, converted, credits earned.ReadEnd client
get_platform_changelogLists latest platform updates (changelog).ReadEnd client
get_showcase_contentGenerated content of a client's showcase page in a given language.ReadPartner
get_showcase_statusClient showcase page status: enabled, slug, public URL.ReadPartner
list_showcase_clientsLists all clients with an active showcase page.ReadPartner
regenerate_showcaseRegenerates showcase page content via AI (article, FAQ, meta).WritePartner
toggle_showcaseEnables or disables a client's showcase page. Generates slug if needed.WritePartner

Data quality alerts 3

Tool Description Mode Access
get_audit_alertsRetrieves data quality alerts for an audit. Returns non-dismissed alerts, sorted by severity (error > warning > info).ReadEnd client
get_client_alertsRetrieves quality alerts from a client's last completed audit. Returns non-dismissed alerts, sorted by severity, limited to 50.ReadEnd client
dismiss_alertDismiss (hide) a data quality alert. Marks alert as dismissed with timestamp.WriteEnd client

Questionnaires 4

Tool Description Mode Access
compare_questionnaire_responsesCompares responses between two audits for the same client.ReadEnd client
get_custom_questionnairesLists a client's custom questionnaires.ReadEnd client
get_questionnaire_completion_statusChecks which questionnaire sections are completed and which are pending.ReadEnd client
get_questionnaire_responsesClient questionnaire responses (optionally for a specific audit).ReadEnd client

Agency 9

Tool Description Mode Access
get_agency_agentsLists all agency agents with role, credits, and client count.ReadAgency
get_agency_clientsLists all clients across all agency agents.ReadAgency
get_agency_credit_poolAgency credit pool details with allocation per agent.ReadAgency
get_agency_dashboardAgency overview: agents, clients, credit pool, recent activity.ReadAgency
get_agency_whitelabelAgency white-label config (logo, colors, business name, contact details, website). All wl_* columns from agencies table.ReadAgency
allocate_agency_creditsAllocates credits from pool to a specific agent.WriteAgency
invite_agency_agentInvites a new agent to join the agency by email.WriteAgency
remove_agency_agentRemoves an agent from the agency (disables their membership).WriteAgency
transfer_agency_clientTransfers a client from one agent to another within the agency.WriteAgency

Analytics & dashboards 7

Tool Description Mode Access
export_client_full_report_dataComplete export of all client audit data in structured JSON.ReadEnd client
get_client_360Complete 360° client view in a single call: profile, contacts, competitors, full audits (latest + previous + history), model breakdown, progress summary, website analysis, enriched action plans, GEO modules (SSR, entity, mention, citation, STS), data alerts. Financial data (credits, subscription, PDF config) visible only to agents and admins.ReadEnd client
get_partner_dashboardPartner dashboard: client portfolio, aggregated scores, audits, credits consumed, top/bottom clients.ReadPartner
get_score_distributionAudit score statistical distribution: histogram, percentiles, mean, median.ReadPartner
get_sector_benchmarksAI sector benchmarks: average/median/percentile metrics by industry sector and AI model. Useful for comparing a client to their sector.ReadPartner
get_visibility_leaderboardClient ranking by visibility score (latest audit). Useful for partners.ReadPartner
search_everythingGlobal cross-table search (clients, audits, contacts, competitors, prompts). Categorized results.ReadPartner

Blog & glossary 5

Tool Description Mode Access
get_blog_article_detailComplete article content with translations, FAQs, and ratings.ReadEnd client
get_blog_article_faqsBlog article FAQs in a specific language.ReadEnd client
get_blog_articlesLists blog articles with filtering.ReadEnd client
get_glossary_term_detailComplete glossary term with translations and related terms.ReadEnd client
get_glossary_termsLists glossary terms with filtering.ReadEnd client

Users 2

Tool Description Mode Access
get_my_profileLogged-in user profile (safe subset of fields).ReadPartner
get_partner_clients_summaryConnected partner's client summary: latest audit, visibility score, action plans.ReadPartner

System & security 1

Tool Description Mode Access
get_user_activity_logMy own activity logs. Each user can view their action history on the platform.ReadPartner

Your MCP key & access 5

Tool Description Mode Access
get_my_access_requestsLists user's tool access requests.ReadEnd client
get_my_mcp_usageLogged-in user's MCP usage statistics.ReadEnd client
get_my_permissionsLists tools authorized on the API key being used.ReadEnd client
list_available_toolsLists MCP tools accessible according to your role. Each role only sees its authorized tools.ReadEnd client
request_tool_accessSubmits a tool access request. Creates a request pending admin validation.WriteEnd client

Give MCP access to your clients

As a partner with an Agence+ plan, you can create scoped client keys that give your end clients direct access to their own data only via their favorite AI assistant. This is perfect for agencies who want to empower their clients with self-service AI-powered analytics.

How it works: A client-scoped key is linked to a single client. The end user can only see data for that specific client — audits, scores, action plans, reports, etc. They cannot see your other clients or any partner-level data.

Why give MCP access to your clients?

Giving your clients their own MCP key is much more than a technical feature — it’s a strategic lever that transforms your audit service and strengthens your client relationships. Here’s why it matters:

1. Increase the perceived value of your audits

When you deliver an audit report, the value is often perceived as a one-time deliverable. By giving your client an MCP key, you turn that static report into a living, interactive experience. Your client can ask their AI assistant questions like:

  • "What’s my current visibility score and how has it evolved?"
  • "Which AI model mentions me the most?"
  • "Show me the scenarios where my competitors rank higher than me"
  • "What actions should I prioritize to improve?"

This makes your audit a permanent resource, not just a PDF they read once and file away. The client keeps engaging with the data, which reinforces the value of your work.

2. Boost client retention and recurring revenue

Once your client is actively using MCP to check their scores, compare audits and track their progress, they become engaged in the process. An engaged client is far more likely to:

  • Order follow-up audits — They can see their evolution and want to track improvements
  • Stay subscribed longer — The data becomes part of their routine monitoring
  • Recommend your service — They can demonstrate the value to their own colleagues

3. Reduce your support workload

Without MCP access, every time your client wants a data point, they have to email you or schedule a call. "What was my score on Perplexity last month?" "Can you send me the comparison between my last two audits?"

With MCP, your client gets instant, self-service answers 24/7. They just ask their AI assistant. This frees up your time for higher-value work like strategy and new client acquisition.

4. Build trust through transparency

Giving your client direct access to their raw audit data shows confidence in your work. There’s no "black box" — they can explore results, check individual model responses, and verify scores themselves. This transparency builds trust and positions you as a credible expert.

5. Differentiate your service from competitors

Most agencies deliver a PDF and move on. You deliver a PDF plus a live AI-powered interface where the client can explore their data anytime. This is a premium differentiator that justifies higher pricing and sets you apart from basic audit providers.

Real-world example: Imagine your client is in a meeting with their marketing team. Instead of searching for your last email with the PDF attached, they open Claude or ChatGPT and ask: "Show me our AI visibility evolution over the last 3 months with scores by model." In seconds, they have the answer — and they remember who gave them that power: you.

Step 1 — Create a client-scoped key

1

Go to API & Integrations and click "New key".

2

In Access type, select "End client (single client, read only)".

3

Select the client you want to give access to from the dropdown.

4

Click "Create key". The key will be generated and can be viewed again later (unlike partner keys, client keys are stored and can be revealed anytime via the button).

Step 2 — Send the configuration to your client

Share the following information with your client so they can configure their AI assistant:

Information to share:

  • Server URL: https://mcp.ailabsaudit.com/mcp
  • API key: the key you created (starts with aila_)
  • Transport: streamable-http

Here is a ready-to-send message template you can copy and send to your client:

Hello,

You now have access to your AI visibility data directly from your AI assistant (Claude, ChatGPT, Gemini...).

Here’s how to set it up:

1. Open your AI tool settings (e.g. Claude Desktop > Settings > MCP)
2. Add a new MCP server with these details:
   - Server URL: https://mcp.ailabsaudit.com/mcp
   - API Key: YOUR_KEY_HERE
   - Transport: streamable-http

3. Once connected, you can ask things like:
   - "Show me my latest audit results"
   - "What are my visibility scores on ChatGPT and Claude?"
   - "Show me my action plan"
   - "Compare my last two audits"

Your key gives you read-only access to your own data.
For any questions, feel free to reach out!

Step 3 — What your client can do

With a client-scoped key, your end client has access to 47 tools including:

  • Audits: view results, scores, trends, model comparisons
  • Action plans: track progress, browse recommendations
  • Reports: list and download PDF reports
  • Credits: check balance, view history
  • Profile: view their own profile and activity
Important: Client keys are read-only and scoped to a single client. They cannot list other clients, access partner-level analytics, or perform any administrative actions.

Managing client keys

  • View key: Click the button next to the key to reveal it anytime
  • Regenerate: Click to revoke the old key and create a new one
  • Revoke: Click to permanently deactivate the key — your client will lose access immediately
Tip: You can create multiple keys for the same client (e.g. one for their marketing team and one for their GEO team). Each key works independently.

Frequently asked questions

Is my data secure?

Yes. Your API key gives access only to your own data. A standard user can only see their clients and audits. A partner sees those in their portfolio. Only an admin has access to everything. All communications are encrypted via HTTPS.

Does it cost credits?

No. MCP access is included in your subscription. Only actions that normally consume credits (running an audit, generating a PDF) will also consume them via MCP.

Does it work on mobile?

It depends on the app. Claude on iOS/Android can use already configured MCP servers. ChatGPT on mobile does not yet support MCP Apps. For Gemini, you need to use the CLI (computer only).

Can I revoke my key at any time?

Yes. Go to API & Integrations, and click the delete button next to the relevant key. Access will be cut off immediately.

My AI can’t find the AILabsAudit tools

Check that:

  • Your API key is valid and has not been revoked
  • The server URL is correct: https://mcp.ailabsaudit.com/mcp
  • Your AI client has been restarted after configuration
  • You explicitly mention "AILabsAudit" in your message so the AI knows which tool to use

How many tools are available?

The MCP server exposes over 235 tools across categories: clients, audits, scores, reports, credits, subscriptions, scenarios, AI configuration, newsletter, blog, glossary, users, partners, system and analytics.

Ready to connect your AI?

Create your API key in 30 seconds and start using AILabsAudit from your favorite AI assistant.

Create my API key

Ready to audit your AI visibility?

Create your free account and receive 500 bonus credits.

Create free account