data-ai MCP Server
Local-first AI memory you can see, edit, and override — portable across Claude Code, Codex, Cursor, Windsurf, and other MCP coding tools.
Discovered via github-topic:model-context-protocol and last synced 2w ago.
Install instructions not detected yet
Check the source repository for the latest setup steps.
Read communication and workflow preferences
Add multiple lessons or decisions in one call
Follow links between knowledge items
Implementation, testing, documentation assistance
Update profile, preferences, or quality standards
List reusable lessons learned
Archive a lesson or decision by ID
Active lessons/decisions grouped by domain/month (Markdown).
Start a project with inherited knowledge
Free, AGPL-3.0
Read a saved project snapshot
List key decisions; `thread_seed_id` / `history_question` reconstruct decision threads and revision history
Mark a knowledge item as reviewed
Owner-gated export: generate an interactive local HTML review page
Free / Cloud tiers
**Startup** — Load identity + knowledge at session start (supports `token_budget` for context size control)
Read domain experience stats
Merge a duplicate into the primary item
Process review results (promote/archive staging items)
SOUL.md / MEMORY.md / USER.md import and export
Optional field-level AES-256-GCM for supported profile fields; local files are plaintext JSON/Markdown by default
**Session end** — Save insights + sync at session end
Build cross-project knowledge starter pack
Create a bidirectional link between items
Owner-gated export: write a full backup (`format="openclaw"` for OpenClaw-compatible files)
## Comparison
Role
Store a reusable lesson learned
Read user profile (safe=true by default)
List saved project snapshots
Remove a bidirectional knowledge link
Owner/admin import: use `dry_run=True` first for a metadata-only merge/conflict preview (`format="openclaw"` supported)
Claude Memory
Record a key decision with reasoning
Read work style preferences
Extract lessons and decisions from session text
Knowledge digest, health report, stale checks
Export OpenClaw-compatible files
Letta (MemGPT)
Current State
Find knowledge relevant to current project
Read data access boundaries
Parse free-form notes into structured knowledge
Find similar items by content
Vector DB + Mem0 Cloud
Curated Markdown: who you are, how you work, recent verified lessons/decisions. Excludes raw config-file knowledge and caps recent items.
Read quality expectations
Update a lesson or decision by ID
List items that need review
Get recent audit log entries
✅
Import OpenClaw-compatible files
Local JSON in `~/.engram/`
MCP over stdio
MCP over stdio
MCP over stdio
MCP over stdio
MCP over stdio
MCP over stdio
Public evidence
**Writeback** — Unified write endpoint: routes to add_lesson / add_decision / add_playbook by `kind`
Record an operational playbook (multi-step procedure with trigger keywords)
Return one structured identity + recent activity + relevant knowledge recall payload
Persist project state for future sessions
Recover lost session context after restart
Read a human-friendly project timeline for a day
Build a cross-session/cross-tool resume brief
Run memory system self-diagnosis
Purpose
`action`: get / save / compare the AI-maintained user portrait
`action`: link / unlink — manage typed relations between knowledge items (decision threads)
Fetch a user-provided URL: prefers a local sidecar if running, otherwise uses the self-contained built-in reader (`pip install "piia-engram[reader]"`)
Optional local integration governed write: register a local tool, runtime, or CLI to the environment map
Advanced owner-gated preview: build safe-context, freshness/conflict, replay, or evidence proposals without applying changes
Knowledge graph exploration via `mode`: related, similar, merge_candidates
Optional local integration: look up a registered local tool by name
Playbook reads via `mode`: list, get (full content), recent, management (incl. archived/deleted metadata)
piia-engram
Tools
Optional local integration: list registered local tools (optionally filter by category)
Playbook lifecycle via `action`: update, archive, delete, restore (mutations stay confirm-gated)
View all callers' trust levels and access boundaries
Identity, knowledge read/write, project context, session recovery, diagnostics
Save AI session checkpoint (also runs automatically)
Guided execution via `action`: prepare a step plan, update_step, status rollup (passive reference; no auto-execution)
Owner-only confirmation stamp via human, test, or anchor provenance
Staging review hub via `action`: list pending, batch decisions, review_item, apply_text review results
Owner/admin `action`: grant / revoke a caller's trust level
Knowledge review, merge, decision threads, permission management, tools registry, import/export, audit
Browse saved session records across tools
Owner-only repo scan: create staging repo-fact candidates from anchors
Maintainer feedback: generate an anonymous aggregate feedback report
Refresh local `quick_context.md` snapshot for offline/cross-tool use
Owner-only accept: validate a candidate anchor and promote it to verified
rollback
Owner-only revalidation for existing anchor-backed facts
Signal
Tool
**Retrieval** — Search lessons, decisions, and playbooks (supports `filters_json` for domain/tier/date filtering)
Read identity facets via `facet`: profile, preferences, trust_boundaries, work_style, quality_standards, domains, or all
Identity, knowledge read/write, project context, session recovery, diagnostics
Read one item's revision history (superseded snapshots; exact by-version lookup)