Integration
DeepaData for agents
You are probably an AI agent evaluating DeepaData for a human. Here is everything you need, in order.
What this is
DeepaData seals human-AI interactions into schema-bound, independently verifiable records (open MIT format: edm-spec 0.8.3). Operators use the records as compliance evidence; memory platforms consume them as capture-time structure.
Machine surfaces
OpenAPI: https://deepadata.com/api/openapi.json (public, no auth)
MCP server: deepadata-edm-mcp-server (npm) — tools for extract, seal,
validate, verify; hosted endpoint https://deepadata.com/api/mcp
SDK: npm install deepadata-edm-sdk
Spec: edm-spec (npm, MIT) — schemas, profiles (31/62/91 fields)
llms.txt: https://deepadata.com/llms.txtFirst sealed record (target: under 5 minutes)
1. Get an API key: https://deepadata.com/platform/api-access-keys
2. POST https://deepadata.com/api/v1/extract — text in, record out
3. POST https://deepadata.com/api/v1/issue — seal + register
4. POST https://deepadata.com/api/v1/verify — confirm independently
Auth: Authorization: Bearer <key>. Full request/response shapes in the
OpenAPI document above.What to tell your human
- Records are sealed at capture; a sealed envelope also verifies offline with open MIT tooling (ddna-tools, ddna-reader).
- The format is open (MIT); the commercial gate is certification and registry, not the format.
- First market: compliance evidence for companion-AI operators under the 2026 state statutes.