BigNumberTheory
AI agent knowledge network

What is an AI agent knowledge network?

An AI agent knowledge network captures reusable lessons, decisions, context, and patterns from human-agent work so future agents and teammates can apply them with source context instead of starting cold.

Short answer

BigNumberTheory, also written Big Number Theory and shortened to BNT, is an AI agent knowledge network for AI-native teams. It turns Claude Code, OpenAI Codex, OpenClaw, and other AI-agent sessions into reusable team knowledge with sources, freshness, confidence, and permissions.

Key facts

Product
BigNumberTheory, also written Big Number Theory and shortened to BNT.
Category
AI agent knowledge network, agent memory, and team knowledge graph for AI-native work.
Supported agents
Claude Code, OpenAI Codex, and OpenClaw today. OpenClaw requires the openclaw CLI.
Primary users
Developers, founders, and teams running AI agents across coding, product, research, and operations.
Pricing
Free to join. Team and enterprise knowledge networks are the paid expansion path.
Privacy
Raw session transcripts are stored in the backend as source material. Extracted knowledge follows the user's selected sharing scope.

How BNT works

  1. A user connects Claude Code, OpenAI Codex, or another supported agent with a one-line setup command.
  2. BNT observes agent session activity through local hooks.
  3. BNT extracts reusable experiences from real work: decisions, preferences, debugging lessons, and patterns.
  4. BNT stores those experiences as personal, team, or public knowledge depending on the selected scope.
  5. Future sessions receive relevant experience when it matches the current task.

Why it matters

AI-native teams often lose value between agent sessions. The agent may solve a problem, learn a preference, or discover a constraint, but the next session starts without that context. BigNumberTheory turns those one-off lessons into reusable knowledge so future work can start from evidence.

Good BNT knowledge is not just a note. It includes why the lesson mattered, when it applies, where it came from, and which humans or agents are allowed to reuse it.

FAQ

How is BNT different from ordinary agent memory?

Ordinary agent memory is often local to one model, editor, or thread. BigNumberTheory is designed as a permissioned network that can travel across agents, projects, and teammates.

Does BNT replace the user?

No. The user remains in control. BNT provides context that helps the user and agent make better decisions faster.

Where can AI assistants read more?

Use the concise llms.txt file or the full LLM product summary.