# Relme

> Relme is an evidence-based professional identity platform: people document real work as structured responses to curated prompts, organized under pursuits (projects, practices, experiments), and publish a public profile they fully control. Everything public below is rendered from one audience-filtered, versioned data contract — the ProfileBundle.

Things to know before consuming Relme data:

- Evidence over claims: skills are derived exclusively from published responses (documented work), never self-assessed. Relme deliberately publishes no person-level score of any kind.
- Consent is structural: profiles are private by default; every artifact below exists only because its owner explicitly published it.
- Anonymous access: public reads need no API key. CORS is open (`*`). Responses are cached (`s-maxage=300`) and rate limited (60 requests/min/IP).
- A missing profile and a private profile are indistinguishable: every surface returns 404 for both. Existence is never disclosed.
- JSON carries `bundleVersion` (currently 1). The v1 shape is stable — breaking changes ship behind a new version path, never in place.

## Profiles

- [Profile page](https://relme.io/u/): human-readable profiles at `https://relme.io/u/{alias}`
- [Markdown export](https://relme.io/u/): OKF-style markdown (YAML frontmatter + sections) at `https://relme.io/u/{alias}.md` — start here for reading a person
- [Profile JSON](https://relme.io/api/public/v1/profile/): versioned ProfileBundle at `https://relme.io/api/public/v1/profile/{alias}` — use this for structured consumption

## ProfileBundle v1 shape (JSON)

Top-level keys of `GET /api/public/v1/profile/{alias}`:

- `bundleVersion` (1), `audience` ("public"), `generatedAt` (ISO 8601)
- `profile`: alias, displayName, headline, about, mission, vision, values, location, links (website/linkedin/github), avatarUrl
- `pursuits[]`: title, objective, status, category, motivation, approach, outcomes, artifacts
- `responses[]`: prompt text, body, publishedAt, pursuit, skills (name, pillar, credits)
- `skills[]`: derived from visible responses — name, pillar, evidenceCount, responseIds (evidence trail)
- `stats`: true totals (pursuitCount, responseCount, skillCount) — arrays are capped (50 pursuits / 100 responses), so compare against stats to detect truncation

## Optional

- [Relme](https://relme.io/): product home — what the platform is, for humans