Configure Faraday — the customer-prediction and audience-deployment platform for consumer brands — directly from Cursor.
This plugin connects Cursor's agent to Faraday's hosted Configuration MCP server at https://mcp.faraday.ai/mcp, plus a set of Cursor skills and rules that teach the agent how to use the tools effectively.
- Predict customer behavior — propensity, lookalike, lead scoring, churn, LTV, repeat purchase, upgrade, donation, attendance.
- Discover the natural shape of a customer base — data-driven personas, clustering, customer typology.
- Recommend products, content, offers, or locations personalized per individual.
- Size and visualize addressable market across a geography (TAM / opportunity heatmaps).
- Deploy per-person predictions, segments, and recommendations to Meta / Google / TikTok, Mailchimp / Klaviyo, HubSpot / Salesforce / mParticle, data warehouses, S3 / GCS / SFTP files, the Faraday Lookup API, downloadable CSVs, or a custom webhook.
- Enrich a known person with 1,500+ built-in consumer attributes — demographics, income, homeownership, life events, household composition, geography, interests.
Install from the Cursor Marketplace, or clone this repository into ~/.cursor/plugins/local/ for local development.
On first use, Cursor will walk you through Faraday's OAuth flow. You'll need a Faraday account (sign up); the OAuth flow grants the plugin scoped access to that account's resources. Machine-to-machine (client_credentials) is also supported for headless / CI use.
| Component | Path | Purpose |
|---|---|---|
| MCP server reference | mcp.json |
Tells Cursor to connect to https://mcp.faraday.ai/mcp |
| Tool-usage rule | rules/faraday-tool-usage.mdc |
Background on Faraday concepts and the resource chain |
| Account-configuration skill | skills/configure-faraday-account/SKILL.md |
End-to-end walkthrough: Connection → Dataset → Cohort → Outcome / Persona Set / Recommender → Scope → Target |
| Predict-customer-behavior skill | skills/predict-customer-behavior/SKILL.md |
Focused recipe for propensity / lookalike / churn / LTV |
The Faraday MCP server itself is not shipped here — it's hosted by Faraday and the source lives in the fdy monorepo at src/sojourner/src/mcp.
The MCP exposes roughly 100 tools mapped onto Faraday's existing Configuration API. The headline categories:
- Accounts & billing —
get_current_account,list_accounts,create_account,get_account_billing,get_account_usage,get_usage_stats, … - Connections —
list_connections,create_connection,rotate_connection_credentials, … - Uploads —
list_uploads,create_upload,get_upload - Datasets —
list_datasets,create_dataset,get_dataset_ingress_logs, … - Streams —
list_streams,create_stream,get_stream_analysis, … - Cohorts —
list_cohorts,create_cohort,get_cohort_membership, … - Outcomes —
list_outcomes,create_outcome,get_outcome_analysis,get_outcome_report, … - Persona Sets —
list_persona_sets,create_persona_set,get_persona_set_dimensions,update_persona, … - Recommenders —
list_recommenders,create_recommender,get_recommender_analysis, … - Scopes —
list_scopes,create_scope,get_scope_analysis,get_scope_efficacy, payload + population inspectors, … - Targets —
list_targets,create_target,get_target_analysis,download_target,list_scope_targets, … - Traits —
list_traits,create_trait,get_trait_dimensions, … - Attributes (next-gen built-in consumer data) —
list_attributes,create_attribute, … - Places —
list_places,create_place, … - Market Opportunity Analyses —
list_market_opportunity_analyses,create_market_opportunity_analysis, … - Webhooks —
list_webhook_endpoints,create_webhook_endpoint, … - Graph —
get_graph(full dependency view across all resources) - Shared lifecycle —
archive_resource,unarchive_resource,force_update_resource,delete_resource
Every list/get tool accepts a JMESPath argument; use it to project just what you need and keep the agent's context window small.
The Faraday MCP is OAuth 2.1 with PKCE and dynamic client registration via CIMD (Client ID Metadata Documents). Cursor handles the flow automatically.
MIT — see LICENSE.
- Faraday docs: https://faraday.ai/docs
- Faraday support: [email protected]
- Plugin issues: https://github.com/faradayio/faraday-cursor-plugin/issues