nLight.fit documentation
Full-fidelity reference for the platform: how the tracking ledger, AI coach, voice agent and health integrations work, and how to reach your own data over the Model Context Protocol.
Reading this as a model or an agent? Every page is available as raw Markdown at /docs/<slug>.md. There is an index at /llms.txt and the entire corpus as one document at /llms-full.txt.
Start
- OverviewWhat nLight.fit is - a daily commitment ledger with an AI coach that has full-resolution memory of your history, plus WHOOP, bloodwork and genetics.
- QuickstartFrom a new nLight.fit account to a coached conversation grounded in your own data, in about ten minutes.
- ConceptsThe vocabulary nLight.fit uses - ledger, metric registry, learnings, Mission Statement, correlation tiers, feature store and more.
Platform
- The ledgerHow daily tracking works in nLight.fit - columns, custom metrics, check-in frequency, immediate persistence and CSV import/export.
- Metric referenceEvery metric nLight.fit tracks: 59 canonical fields with types, units and analytical roles.
- AI coachHow the nLight.fit coach decides what to read, how hard to think, and how it grounds answers in your actual record.
- Voice agentReal-time spoken conversation with the nLight.fit coach, with modes for open chat, hands-free data entry and guided onboarding.
- Mission StatementOne goal, three commitments - a structured declaration of intent that nLight.fit measures progress and causality against.
- AnalyticsWhat nLight.fit computes in the background - correlations with statistical controls, personal records, readiness, goal ETA, changepoints and narrative summaries.
- Progress picturesPhoto tracking in nLight.fit, stored against your record alongside the numbers.
Integrations
- WHOOPConnect WHOOP to nLight.fit - OAuth, sync schedule, webhooks, and which recovery, sleep and strain metrics become columns.
- BloodworkUpload lab report PDFs to nLight.fit - biomarkers are extracted into a structured record with trends, reference ranges and behaviour correlations.
- Methylation geneticsUpload a methylation genetic test to nLight.fit - a one-time profile that explains mechanisms and predicts which biomarkers to watch.
Developers & Agents
- MCP servernLight.fit exposes your health record to external AI clients over the Model Context Protocol, as a stateless HTTP JSON-RPC endpoint with read-only tools.
- MCP tool referenceEvery tool exposed by the nLight.fit MCP server: 16 read-only tools with full input schemas.
- AuthenticationHow nLight.fit API tokens work - format, scopes, storage, revocation, and why they are not session tokens.
- Client setupConnect Claude Code, Claude Desktop, Cursor or a custom client to the nLight.fit MCP server.
- Token managementMinting, listing and revoking nLight.fit API tokens with the repo-local management scripts.
- Errors and limitsStatus codes, JSON-RPC error codes, rate limits and protocol versions for the nLight.fit MCP endpoint.
Trust
- Security and privacyHow nLight.fit protects your data - encryption, authentication, AI safety, third-party handling, and an honest account of what is not protected.
- Data modelWhat nLight.fit stores - collections, the shape of a tracked day, derived artifacts, and how the pieces relate.
- ChangelogUser-visible changes to the nLight.fit platform, newest first.