Conductor MCP
Query your Conductor AEO and SEO data in plain language — brand mentions in AI answers, citations, keyword rankings and account configuration. Read-only, gated behind a paid Conductor subscription, and metered in tool calls.
Verified connector
Listed by Anthropic as a partner connector in its Connectors Directory.
Connection checked by Agentman on .
Anthropic states this reflects the level of review a connector received, not a security audit.
Connect Conductor MCP via MCP
https://mcp-universal.conductor.com/mcp/v3Works in any MCP-compatible client. In Agentman Studio it is one click — no config file to edit.
Conductor MCP Tools & Capabilities (4)
Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.
Limits
- A paid Conductor subscription is required. There is no free tier and no anonymous mode. Conductor states an active Conductor login is needed to authenticate, and the connectors are for customers.
- Tool calls are metered per plan. Each subscription includes an allocation refreshed annually. When it runs out, requests are declined rather than billed — Conductor documents no overage charges.
- A single question can spend several tool calls. Conductor notes an assistant may call
tracked_configs, then a data tool, then drill in. A skill can consume several in one run. - Rate limits are tight. 30 requests per hour per user and 120 per minute system-wide, introduced when the MCP opened to all customers. Increases are considered case by case.
- The directory listing is behind the server. Anthropic's snapshot names four tools; Conductor's docs name five. Conductor attributes the lag to platform app review, from days to months.
- No writes. Conductor states the MCP does not write back into the platform, so this is a reporting surface only.
- Data is a scheduled snapshot, not a live query. Collection runs Daily, Weekly or Monthly per topic. New topics can take five hours to a few days to appear.
- Coverage is bounded by what you pay to track. Conductor generates up to 100 prompts per topic, and topic and prompt counts are limited by your AI Response Credit budget. Untracked prompts return nothing.
- Google AI Overviews responses are incomplete by nature. Conductor documents that AI responses are generated for only 50–70% of prompts there.
- No AI search-volume metric and no aggregated share-of-model number. Conductor declines to publish either, citing the absence of a reliable measurement method.
- Some capabilities are documented as not yet available, including pulling exact citation URLs for every prompt and direct BI-tool integration. Gemini is not a supported client.
- Richer skills are uneven across assistants. Conductor documents multi-step skills as optimised for Claude and ChatGPT, with Copilot support still being explored.
- We could not verify tool behaviour ourselves. The endpoint returns 401 to an anonymous handshake, so no schemas and no safety annotations were readable. The read-only claim is Conductor's documentation, not our observation.
- Prompts are unconfirmed. Anthropic's snapshot lists no prompt names, and the gated endpoint blocked our own
prompts/list, so we cannot say whether the server serves any.
Frequently asked questions
Do you need a paid Conductor subscription to use the Conductor MCP server?
Yes. Conductor states the connectors are available to customers with an active Conductor login, which is required to authenticate. Conductor is an enterprise AEO and SEO platform sold on annual contracts, so there is no free tier to connect. Its own troubleshooting guide names a 403 with a plan message as the sign that MCP access is not enabled on your subscription.
How does Conductor measure whether a brand is cited by AI assistants?
Conductor submits tracked prompts to AI engines and records the answers. It states it uses the official APIs published by each supported engine rather than scraping. A citation is counted when a link to your site appears among the sources behind a response, and repeated links each count separately, while repeated brand mentions in one answer count only once.
Which AI engines does Conductor track for brand mentions and citations?
Conductor lists eight in its AI Search setup guide: ChatGPT Auto, ChatGPT Search, Claude Sonnet, Gemini, Google AI Mode, Grok Auto, Microsoft Copilot and Perplexity. Coverage spans over 160 countries, though city-level tracking is unavailable on Google AI Mode. Fan-out query data is narrower, documented for Gemini, Claude, ChatGPT and Grok only.
How fresh is the AI citation data the Conductor MCP returns?
It depends on the tracking cadence you chose per topic. Conductor offers Daily, Weekly and Monthly collection, and warns that daily tracking substantially increases credit usage. New topics take as little as five hours to appear and sometimes a few days. So the connector reads a scheduled snapshot of AI answers, not answers generated at query time.
Can the Conductor MCP server write or change anything in my Conductor account?
No. Conductor's own FAQ states the MCP is focused on querying and analysing your data and does not write back into the platform. All the tools return metrics or configuration metadata. Anthropic's directory record carries no permissions label at all, so this read-only claim rests on Conductor's documentation rather than on a label we could check.
Why does the Conductor connector in Claude's directory show fewer tools than the docs list?
Because directory listings lag Conductor's server. Conductor explains that each native app and directory connector goes through the AI platform's review process, so newly released tools can take days or months to appear. Anthropic's 2026-08-16 snapshot names four tools; Conductor's Claude setup page names five. A custom connector always runs the newest version.
How many Conductor MCP tool calls does a subscription include?
Conductor allocates a set number per plan, refreshed each contract year, and shows the figure on the Subscription page for account admins. A call counts only when a data tool successfully returns data. Conductor also documents rate limits of 30 requests per hour per user and 120 requests per minute across the whole system, with no overage billing.
Sources
- Conductor MCP introduction — https://www.conductor.com/docs/mcp/what-is-mcp/ (fetched 2026-08-19; this is the
documentationURL Anthropic's directory publishes) · retrieved 2026-08-19 - What is an MCP tool? — https://www.conductor.com/docs/mcp/what-is-an-mcp-tool/ (fetched 2026-08-19) · retrieved 2026-08-19
- What data is available in Conductor's MCP? — https://www.conductor.com/docs/mcp/what-data-is-available-in-conductors-mcp/ (fetched 2026-08-19; per-tool descriptions and filters) · retrieved 2026-08-19
- What can I do with Conductor's MCP? — https://www.conductor.com/docs/mcp/what-can-i-do-with-conductors-mcp/ (fetched 2026-08-19) · retrieved 2026-08-19
- Conductor MCP Quick Start Guide — https://www.conductor.com/docs/mcp/quick-start/ (fetched 2026-08-19) · retrieved 2026-08-19
- Conductor MCP setup for Claude — https://www.conductor.com/docs/mcp/claude/ (fetched 2026-08-19; names five tools and the plan requirements) · retrieved 2026-08-19
- Conductor Custom Connectors — https://www.conductor.com/docs/mcp/custom-connectors/ (fetched 2026-08-19; API-token authentication) · retrieved 2026-08-19
- Conductor MCP FAQs — https://www.conductor.com/docs/mcp/mcp-faqs/ (fetched 2026-08-19; access gating, metering, rate limits, no-write statement, listing lag) · retrieved 2026-08-19
- Conductor AI Search FAQs — https://www.conductor.com/docs/intelligence/ai-search-faqs/ (fetched 2026-08-19; measurement methodology, mention and citation counting, data latency) · retrieved 2026-08-19
- Conductor AI Search Setup — https://www.conductor.com/docs/intelligence/ai-search-setup/ (fetched 2026-08-19; supported engines, tracking cadence, credit forecasting) · retrieved 2026-08-19
www.conductor.com/robots.txtallows all user agents apart from a favicon path, names two sitemaps, and carries noContent-Signalline on any axis (fetched 2026-08-19). The host serves anllms.txt, but it indexes gated PDF marketing assets rather than documentation; the docs sitemap atwww.conductor.com/docs/sitemap.xmlwas the machine index that yielded the 11 MCP pages. Appending.mdto a docs URL returns a 307 redirect rather than Markdown · retrieved 2026-08-19- Live RFC 9728 protected-resource descriptor — https://mcp-universal.conductor.com/.well-known/oauth-protected-resource (fetched 2026-08-19; the path-append form 404s and the path-insert form returns an empty 401) · retrieved 2026-08-19
- Live RFC 8414 authorization-server metadata — https://mcp-universal.conductor.com/.well-known/oauth-authorization-server (fetched 2026-08-19) · retrieved 2026-08-19
- Live anonymous probe — 2026-08-19:
POSTtohttps://mcp-universal.conductor.com/mcp/v3returns HTTP 401 with a barewww-authenticate: Bearerheader, no scope and noresource_metadata. The/mcp/v1,/mcp/v2and bare/mcppaths each return HTTP 401 identically · retrieved 2026-08-19 - Anthropic connector directory — https://claude.ai/directory/cd4afca5-534c-4794-a3fb-d7b81cde7c4e (snapshot 2026-08-16; four tool names, no permissions field, no prompt names) · retrieved 2026-08-16
Use in Agentman
Connect once and your agents call these tools on their own — on a schedule, in a workflow, with nobody at the keyboard.
Open in Agentman StudioServer Info
- Category
- Analytics
- Developer
- Conductor MCP
- Tools
- 4
- Domain
- mcp-universal.conductor.com
Using Claude Desktop or another MCP client? Setup docs — the connection URL above works anywhere.