MCP

Faraday is a customer context platform: it enriches the people and prospects you care about with 1,500+ consumer and identity attributes, resolves identities, builds audiences and personas, trains predictive models, and powers recommendations — then delivers that context wherever you need it.

Faraday's MCP server lets an AI agent configure and operate your Faraday account directly — connecting data, enriching and resolving identities, defining populations, training models, and deploying per-person context to your destinations. Instead of clicking through the dashboard or writing HTTP requests against the API, you describe what you want and the agent provisions the resources for you using the idiomatic Model Context Protocol.

The server URL is:

https://mcp.faraday.ai/mcp

Before you begin

You need an existing Faraday account. The MCP server connects an AI agent to an account you already have — it doesn't create one for you. If you don't have a Faraday account yet, start at faraday.ai.

The MCP server is available on every Faraday account, with no opt-in or setup required — just connect your agent and sign in.

Connecting an agent

Faraday's MCP server uses OAuth — you sign in to Faraday through your browser, and the agent never handles an API key. Any MCP-compatible client works; the exact steps depend on the client.

Claude (Desktop, Code, or claude.ai)

Add a custom connector pointing at https://mcp.faraday.ai/mcp. When you enable it, a browser window opens to sign in to Faraday and authorize the connection. Once authorized, the Faraday tools are available in your conversation.

Cursor and other clients

Most clients accept a remote MCP server URL in their settings. Add https://mcp.faraday.ai/mcp as the server URL and complete the browser sign-in when prompted.

Confirming your connection

After connecting, ask your agent to run the whoami tool. It reports which Faraday identity your session resolved to and lists every tool the server exposes — a quick smoke test that the connection and your identity are working.

What you can do

The server exposes a comprehensive set of tools covering the full Faraday workflow, from raw data to delivered context:

• Accounts & billing — inspect your account, manage subaccounts, and check usage, quotas, and billing.
• Connections — wire up external data environments (Snowflake, BigQuery, Postgres, S3, GCS, Azure Blob, HubSpot, Salesforce, and more) for ingest and delivery.
• Uploads & datasets — upload flat files and map connected or uploaded data into people, per-person traits, and behavioral event streams.
• Cohorts — define populations of people from trait and event predicates.
• Outcomes — train propensity models (churn, conversion, repeat purchase, lead scoring, and more).
• Persona sets — segment a cohort into data-driven personas.
• Recommenders — train personalized product, content, or location recommendation models.
• Scopes & targets — bundle context into a deployable payload and deliver it to ad platforms, ESPs, CRMs/CDPs, data warehouses, files, the Lookup API, or custom webhooks.
• Traits, attributes & places — browse Faraday's 1,500+ built-in consumer and identity attributes, define your own, and create geographic areas for location-based targeting.
• Market opportunity analyses — size and map an addressable market across a geography.
• Webhooks & graph — subscribe to resource lifecycle events and inspect your account's full dependency graph.

The agent can also archive, unarchive, re-run, and delete resources as part of normal lifecycle management.

Switching accounts

If you have access to more than one Faraday account — for example, a parent account and its subaccounts — you don't need to reconnect to move between them. Ask the agent to list your joinable accounts and join a different one; every tool call afterward runs against the newly joined account.

Getting help

Please don't hesitate to reach out to your account team or open a ticket in our support portal if you have any questions or feedback.