# Outlit
> Infrastructure for customer context. Outlit unifies product, billing, and conversation data into a structured customer profile your agents can query before they act.
## What Outlit Does
Outlit is customer context infrastructure with three core components:
**The Complete Customer Profile**
Every source in your stack, resolved into one canonical customer profile. Usage, billing, support, and conversations in one clean record.
**Establish Facts Across Every Customer**
Outlit extracts facts and signals from emails, calls, Slack threads, and support tickets. Unstructured text becomes structured data, attached to the right customer.
**The Customer API**
Query any customer from your CLI, MCP, or API and pipe the results into any agent or workflow.
## Data Model
Outlit builds a Customer Context Graph with four entity types:
- Visitors — anonymous users tracked via browser SDK before identification
- Contacts — identified individuals with email, userId, or device fingerprint
- Accounts — organizations grouping multiple contacts, linked to billing
- Devices — cross-device associations for the same person
**Contact Journey Stages (unidirectional)**
- Discovered — email known, hasn't signed up
- Signed Up — created an account
- Activated — completed onboarding milestone
- Engaged — actively using the product (auto-detected over 30+ day windows)
- Inactive — no activity for 30+ days (returns to Engaged on resumption)
**Account Billing Statuses (independent of journey)**
- None — never had a subscription
- Trialing — active trial
- Paying — active paid subscription
- Churned — had subscription, now cancelled
**Identity Resolution**
Resolves across four identifier types in priority order: Email (highest), External User ID, Fingerprint, Anonymous ID (lowest). Automatic form detection triggers identify() on email field submission.
**Facts Layer**
Machine-generated assertions with temporal validity:
- Behavioral — e.g., "Core feature usage declined 40% over 14 days"
- Extracted — parsed from communications (e.g., competitor mentions in emails)
- Synthesized — cross-source correlations indicating churn risk or expansion opportunity
## Integrations
- Slack: messages, reactions, thread context
- Gmail: email threads, sender history, attachments
- Stripe: revenue, subscriptions, plan changes, churn events
- PostHog: feature usage, session activity, events
- Fireflies.ai: call transcripts, action items, summaries
- Granola: meeting notes, transcripts, summaries, attendees
- Pylon: support tickets, customer issues, conversations
- Google Calendar: meetings, attendees, scheduling patterns
- Supabase Auth: user identity, sign-up source, auth events
- Clerk: user profiles, org membership, login activity
- Cal.com: meeting bookings, scheduling events
- Calendly: meeting bookings, scheduling events
- Browser SDKs: React, Vue, Next.js, Nuxt, Angular, Astro, SvelteKit
- Server SDKs: Node.js, Rust
## MCP Integration
Endpoint: https://mcp.outlit.ai/mcp
Auth: Bearer token via HTTP header
Compatible with: Claude Code, Claude Desktop, Cursor, Codex, any MCP client
**Available MCP Tools**
- outlit_list_customers — list/filter customers by billing status, activity, MRR, type
- outlit_list_users — list/filter contacts by journey stage, activity level, customer scope
- outlit_get_customer — full customer profile with users, revenue, timeline, behavior metrics
- outlit_get_timeline — activity history filtered by timeframe, channels, event types
- outlit_search_customer_context — natural language semantic search across all context
- outlit_get_facts — AI-generated insights with severity and temporal validity
- outlit_query — read-only SQL against analytics tables
- outlit_schema — table schema introspection
**SQL Tables**
- events — raw event log (pageviews, form submissions, backend events)
- customer_dimensions — customer attributes (domain, billing status, MRR, activity)
- user_dimensions — contact attributes (email, journey stage, activity)
- mrr_snapshots — monthly recurring revenue over time
## CLI
```
curl -fsSL https://outlit.ai/install.sh | bash
```
**Commands**
- outlit customers list — filter by --billing-status, --mrr-above/below, --no-activity-in, --search
- outlit customers get <customer> — accepts domain, UUID, or name; --include users/revenue/recentTimeline/behaviorMetrics
- outlit customers timeline <customer> — --channels EMAIL/SLACK/WEB, --event-types, date range
- outlit users list — --journey-stage, --customer-id, activity filters
- outlit facts <customer> — AI-generated insights with --timeframe
- outlit search <query> — semantic search with --customer scope, --top-k
- outlit sql <query> — read-only SQL; --query-file; tables: events, customer_dimensions, user_dimensions, mrr_snapshots
- outlit schema [table] — list tables or describe columns
## Example Queries
- Find churning customers: outlit customers list --billing-status CHURNED
- Get full customer profile: outlit customers get acme.com --include users,revenue,recentTimeline
- Search for competitor mentions: outlit search "mentioned competitor" --customer acme.com
- Check at-risk users: outlit users list --journey-stage AT_RISK
- Revenue analysis: outlit sql "SELECT * FROM mrr_snapshots ORDER BY snapshot_date DESC LIMIT 10"
## Pricing
- Free: $0/mo, no credit card required, CLI + MCP access, unified customer profile, 3 connections, 1,000 API calls/mo
- Builder: $49/mo, everything in Free, 6 connections, Slack bot access, 3,000 API calls/mo, $0.02/call over limit
- Pro: $199/mo, everything in Builder, unlimited connections, 10,000 API calls/mo, proactive signals, priority support + SLA, $0.01/call over limit
- Enterprise: Custom, everything in Pro, custom API call volume, custom agent and workflow buildout, dedicated SLA, SSO, and support
## Links
- Docs: https://docs.outlit.ai
- MCP: https://docs.outlit.ai/ai-integrations/mcp
- Pricing: https://outlit.ai/pricing
- Blog: https://outlit.ai/blog
- Full context: https://outlit.ai/llms-full.txt
- Contact: https://cal.com/josh-earle
## Security
SOC 2 Type I & II certified. ISO 27001 certified.
Data encrypted at rest and in transit. Never used to train models.
Backed by Y Combinator
