What is Brainfork?
Brainfork is a hosted memory layer specifically designed for OpenClaw AI agents that addresses the critical problem of session-based memory loss. Traditional AI agents start each session fresh, forcing users to repeatedly explain project details, previous decisions, and working conventions. This tool stores documents, decisions, and retrieval context in an external memory layer that remains portable and inspectable, ensuring knowledge persists beyond individual sessions.
The solution integrates seamlessly with OpenClaw through a dedicated plugin that automatically syncs memory files like MEMORY.md and memory documents. Before each agent session begins, Brainfork recalls relevant context and injects it into the session, eliminating the need to manually provide background information. This approach maintains knowledge sovereignty by keeping memory data under user control in their own Brainfork server, separate from model vendors.
Features
- Memory Persistence: Stores documents, decisions, and context that survive agent restarts and carry forward between sessions
- Automatic Context Injection: Recalls relevant context before each agent session starts without dumping entire archives into prompts
- Memory Sovereignty: Keeps knowledge in a user-controlled Brainfork server, preserved independently of model vendors
- Decision Logging: Captures and tags significant agent decisions, building an auditable history of project choices
- File Synchronization: Automatically syncs MEMORY.md and memory documents with deterministic hashes, updates, and cleanup
Use Cases
- Eliminating repetitive explanations to AI agents at the start of each session
- Maintaining project continuity across multiple AI agent sessions
- Building an auditable history of AI agent decisions and project choices
- Keeping working memory synchronized across different AI tools and platforms
- Preserving knowledge sovereignty when using multiple AI model providers
FAQs
-
What AI tools are compatible with Brainfork?
Brainfork is built specifically for OpenClaw and uses the OpenClaw plugin system to integrate automatically with agent sessions. Standard MCP tools like Claude, Cursor, and ChatGPT can also connect to the knowledge base. -
How does Brainfork handle data security?
Your data stays in a Brainfork server you control. The company does not use your data to train models, and you can export or delete your memory at any time from your dashboard. -
Can I use Brainfork with multiple AI agents simultaneously?
Yes, each agent workspace syncs its own memory independently. You can have multiple agents all writing to and reading from the same Brainfork server. -
What happens if I exceed my monthly usage limits?
Overage charges apply at €1 per 10k extra requests and €1 per 250 extra indexed documents. You can monitor usage in your dashboard before any overage charges occur.
