0 memories shaped into a cleaner operational layer across 60 traces and 6 workflows.
OpenClaw memory already lives in workspace files like MEMORY.md and memory/YYYY-MM-DD.md. OpenTrust starts from that authored truth instead of inventing a fake replacement story.
Ingest the existing memory files, then add retention classes, provenance, review workflows, historical timelines, and exportable backup paths.
Guide users toward daily summaries, long-term profile memory, project memory, and meaningful life moments that deserve durable storage beyond chat history.
Track freshness, search quality, and coverage while preparing a first-class OpenClaw plugin with operator surfaces, backups, and future CRM-style relationship memory.
OpenClaw already captures memory in workspace Markdown. OpenTrust should ingest that source first, then add structure, provenance, recommendations, backups, and operator tooling around it.
Import the real OpenClaw memory source today: MEMORY.md, memory/YYYY-MM-DD.md, and related workspace Markdown files.
Generate intelligent prompts and drafts for daily summaries, unresolved threads, and meaningful moments worth preserving.
Keep OpenClaw’s searchable memory feel while adding stronger structure, indexing, and operator review.
Connect long-term memory back to workspace files, traces, artifacts, workflows, and future writeback sources.
Recommend git-backed history, encrypted cloud backups, export bundles, and dual-store strategies for durable memory archives.
Plan for relationship memory, contacts, interactions, and follow-up queues in docs first before runtime implementation.
OpenClaw already has real memory. OpenTrust doesn't replace it — it makes it reviewable, traceable, and durable.
Workspace Markdown is already the canonical authored memory source.
Same source, now with structure, review, provenance, and durability.
MEMORY.md and daily notes already work. Agents read and write workspace memory files with memory_search recall.
OpenTrust ingests those same files. No replacement, no fake abstraction — just operational structure on top of authored truth.
Memory capture exists, but no dedicated review queue, retention lifecycle, or structured approval workflow.
Every memory candidate gets review status, retention class, provenance links, and operator visibility.
Memory truth exists in files, but backup strategy, timeline surfaces, and operational health are mostly implied.
Calendar views, health monitoring, export flows, and backup-aware storage recommendations become first-class.
Users can write summaries manually, but no built-in guidance about what deserves durable capture.
Daily summary drafts, meaningful moment detection, and promotion suggestions based on actual activity.
OpenTrust should begin by respecting the memory system OpenClaw already uses today, then layer operational structure and user-facing capabilities on top.
Markdown stays yours and human-readable. OpenTrust adds structure, search, exports, and backups around it — never instead of it.
Four steps. Start from what OpenClaw already stores today, then layer in structure, recommendations, and full plugin-grade operations.
Import MEMORY.md, memory/YYYY-MM-DD.md, and optional session-memory snapshots from the OpenClaw workspace.
OpenTrust doesn't just store memories — it intelligently suggests what deserves durable capture, when to summarize your day, and when a moment looks historically meaningful.
Docs-only for now: people, interactions, relationship memory, and follow-up queues with explicit privacy guardrails.
Start from the memory OpenClaw already stores today, then layer in review, provenance, backups, summaries, and future plugin-grade memory operations.
# Clone and install
git clone https://github.com/OpenKnots/OpenTrust
cd OpenTrust
pnpm install
# Initialize database
pnpm run db:init
# Ingest OpenClaw evidence + memory files
pnpm run ingest:openclaw
# Launch dashboard
pnpm dev