Console

Loading…

Account

Get started: about five minutes to a confirmed outcome

  • 1  Create a sandbox key with the Sandbox key button below. Free on every plan, it does not consume a production key slot, and nothing it writes can reach live data or a real score. Shown once; keep it server-side.
  • 2  Run one command. It seeds your sandbox with synthetic subjects, prints the score the real formula gave each of them, and reads one back:
    npx @credda/cli quickstart
    No CLI? One HTTP call does the seeding: curl -X POST https://api.credda.io/api/v1/test/seed -H "Authorization: Bearer YOUR_TEST_KEY"
  • 3  Read a score you did not have to invent. Every score endpoint now returns something:
    curl https://api.credda.io/api/v1/users/sbx_reliable_courier/score/explain   -H "Authorization: Bearer YOUR_TEST_KEY"
  • 4  Earn one, the part that matters. Reading a seeded score proves the wiring; a counterparty-confirmed outcome is the product. Propose one, then play the counterparty yourself (you hold the one-time token while you integrate) and watch a verified event land:
    curl -X POST https://api.credda.io/api/v1/confirmations \
      -H "Authorization: Bearer YOUR_TEST_KEY" -H "Content-Type: application/json" \
      -d '{"userId":"sbx_confirmation_demo","eventType":"CONTRACT_FULFILLED",
           "counterpartyRef":"ops@client.example","description":"Kitchen refit, completed 12 March."}'
    
    # the response carries confirmationToken (shown once) + confirmUrl. Then, with NO key:
    curl -X POST https://api.credda.io/api/v1/confirmations/CONFIRMATION_ID/respond \
      -H "Content-Type: application/json" -d '{"token":"THE_TOKEN","decision":"confirm"}'
    Free on every plan and no events:write scope needed: creating a request writes nothing, and the event is recorded only because a party other than the subject confirmed it. Full walkthrough: the counterparty loop.
  • 5  Then wire it up for real: the quickstart guides, the partner arc, or install a client: @credda/js (typed SDK), @credda/cli, @credda/mcp-server (AI agents).

Done experimenting? DELETE /api/v1/test/data wipes the sandbox. It is the only deletion path in the product, and every filter it uses asserts test mode, so the live ledger stays append-only.

Plan & usage

Plan
API calls
Rate limit
Active keys

Official tier pricing is on pricing; self-serve checkout is coming, and nothing is charged through this console today. Your account itself is managed on credda.io.

API keys

LabelModeKeyScopesCreatedLast usedStatus
Loading…

Keys authenticate server-to-server calls. Scopes follow your plan; a key is shown in full only at creation, and revoking one takes effect immediately. A crd_test_ key is free on every plan, keeps its own quota, and sees only sandbox data; it can never read, write or move anything live. A crd_live_ key needs a paid plan.

Your plan

tier

Loading plan…

    Scopes are least-privilege and enforced server-side. Compare tiers on pricing.

    Tools & resources

    Typed SDK

    The @credda/js client: score reads, event writes, webhook verification, offline credential checks.

    npm install @credda/js
    Agent tooling →

    Command line

    The @credda/cli for terminals and CI: lookups, exports, offline verify, key-scoped reads and mints.

    npm install -g @credda/cli
    API reference →

    MCP server

    The @credda/mcp-server lets any MCP-aware agent check or present Credda trust mid-reasoning.

    npm install @credda/mcp-server
    Agents & AI →

    Reference & contract

    Every endpoint, request and response, plus the machine-readable spec.

    Interactive reference → /openapi.json

    Webhooks

    HMAC-signed score.* and dispute.resolved push events, with retries and replay. Manage them from credda.io billing, per key.

    Webhooks & automation →

    Interoperability

    Offline verification, did:web:api.credda.io and the public trust registry, no key required.

    Credential & verification →