Documentation
Memory Platform Integrations
EDM enriches your existing memory stack — one function call, 57 affective fields, +50pp recall lift on the live /v1/activate endpoint.
Controlled retrieval comparison
EDM field routing on the live /v1/activate endpoint: 83.3% (15/18) hit rate on significance-typed queries. Raw vector similarity: 33.3% (6/18). Queries EDM answered exclusively had zero lexical or semantic overlap — they required affective structure to find.
Mem0 + EDM
Mem0 stores and retrieves. EDM encodes what's worth retrieving — significance, arc type, recall triggers, identity thread.
Install
npm install deepadata-mem0-adapter
Usage
import { enrichWithEDM } from 'deepadata-mem0-adapter'
const { edmArtifact } = await enrichWithEDM(text, {
profile: 'essential'
})Zep + EDM
Zep graphs temporal memory. EDM adds significance weight to the edges — emotional_weight, tether_type, recurrence_pattern map directly to graph semantics.
Install
npm install deepadata-zep-adapter
Usage
import { enrichWithEDM } from 'deepadata-zep-adapter'
const { edmArtifact } = await enrichWithEDM(text, {
profile: 'essential'
})LangChain + EDM
LangChain manages memory modules. EDM adds a governed schema — structured significance alongside your existing memory type.
Install
npm install deepadata-langchain-adapter
Usage
import { enrichWithEDM } from 'deepadata-langchain-adapter'
const { edmArtifact } = await enrichWithEDM(text, {
profile: 'essential'
})How it works
- 1.Install the adapter for your platform
- 2.Call
enrichWithEDM(text)alongside your existing memory call - 3.The artifact is structured and ready for significance-aware retrieval
Profiles
| Profile | Fields | Best for |
|---|---|---|
| Essential | 24 | Lightweight capture |
| Extended | 50 | Journaling, companion |
| Full | 57 | Clinical, regulated |