Git-like persistent memory for OpenClaw agents. Your agent remembers everything across sessions — versioned, branched, and mergeable like a repository. Never lose context again. Powered by Qoris AI.
OpenClaw agents are powerful. But by default they forget everything when a session ends. Every conversation starts from zero. Every context you built up disappears.
Qoris Memory gives your OpenClaw agent persistent, versioned, cross-session memory that works like Git — commits, branches, merges, and conflict resolution. Your agent builds up knowledge over time, remembers it across sessions, and shares it across your entire team.
Think of it as GitHub for your agent's brain.
Everything your agent learns, discovers, or is told persists beyond the current session. Next session it picks up exactly where it left off. No re-explaining context. No repeating yourself.
Every memory update is a versioned commit with a timestamp and author. Roll back to any previous state. See the full history of what your agent knows and when it learned it.
Create separate memory branches for different projects, clients, or contexts. Switch between branches the same way you switch Git branches. Your agent operates in the right context for the right task.
When multiple agents or team members update the same memory, Qoris Memory handles conflicts intelligently — surfacing disagreements for human resolution rather than silently overwriting important context.
Memory is workspace-scoped. Every agent in your workspace shares the same knowledge base. One agent learns something — all agents know it. Your AI team operates with a unified brain.
Semantic search across everything your agent knows. Ask questions about your memory — get precise, cited answers grounded in what was actually stored, not hallucinated.
Qoris Memory uses a canonical + vector hybrid architecture:
Canonical Layer — structured facts, entities, relationships
exact-match retrieval, versioned records
Vector Layer — semantic embeddings for fuzzy search
conceptual retrieval across all memories
Conflict Engine — detects contradictions between memories
surfaces them for human resolution
Audit Trail — every memory read and write logged
integrated with Knox governance if installed
Once installed your agent has access to these tools:
Store a new memory with optional tags and metadata.
Retrieve all memories or filter by tag, date, or relevance.
Semantic search across your entire memory and knowledge base.
Update an existing memory with version tracking.
Remove a memory with audit trail entry.
List all documents and files indexed in the knowledge base.
Retrieve the complete content of a knowledge document.
export QORIS_API_KEY="your-api-key-here"
export QORIS_WORKSPACE_ID="your-workspace-id-here"
Add to your OpenClaw configuration:
{
"mcpServers": {
"qoris-memory": {
"url": "https://mcp.qoris.ai/mcp",
"headers": {
"Authorization": "Bearer ${QORIS_API_KEY}",
"X-Workspace-ID": "${QORIS_WORKSPACE_ID}"
}
}
}
}
Start a new OpenClaw session and run:
/memory status
If you have Knox Governance installed alongside Qoris Memory, every memory read and write is automatically logged in the Knox audit trail. Install both for the complete governed enterprise agent stack:
clawhub install knox-governance
clawhub install qoris-memory
Memory is workspace-scoped. Free tier includes up to 1,000 memories and 500MB knowledge storage. Paid plans unlock unlimited memories and storage.
Qoris AI is the trust and governance layer for enterprise AI agents. Knox governs what agents do. Qoris Memory gives them what they know.
NVIDIA Inception Program member. Claude Partner Network member. Patent Pending U.S. 63/907,730. Based in Stamford, CT.
qoris.ai