GrantAi Memory

Give your OpenClaw agents persistent memory that works across sessions. GrantAi uses exact-recall architecture (not RAG/vectors) for sub-second retrieval, 100% local with AES-256 encryption.

Setup

Install GrantAi:

curl -fsSL https://solonai.com/install.sh | bash

Add to mcporter:

mcporter config add grantai ~/.grantai/bin/grantai-mcp
mcporter tools grantai

skill.json Example

{
  "name": "research-assistant",
  "mcpServers": ["grantai"],
  "systemPrompt": "You have persistent memory via GrantAi. Query grantai_infer before searching files."
}

Available Tools

• grantai_infer — Query memory. Use before searching files or re-reading context.
• grantai_teach — Store facts, decisions, and learnings for future sessions.
• grantai_learn — Import files or code into memory.
• grantai_summarize — Save a session summary before ending.
• grantai_project — Track project state across sessions.
• grantai_snippet — Save reusable code patterns.

When to Use Each Tool

Situation	Tool
Starting a session	grantai_infer — check what's already known
User corrects or updates facts	grantai_teach
Onboarding a new codebase	grantai_learn
End of session	grantai_summarize
Tracking feature/sprint state	grantai_project
Reusable code pattern found	grantai_snippet

Multi-Agent Coordination

Agents share memory automatically — a researcher, writer, and reviewer agent can all read and write the same GrantAi brain without extra configuration.

Notes

• 30-day free trial, no credit card needed.
• Docs: https://solonai.com/grantai/integrations/openclaw
• Support: support@solonai.com