Belief-state runtime for AI agents. Track what your agent believes, how confident it is, and why.

BDI reasoning engine: beliefs, intentions, inference rules, working memory

Cognitive infrastructure for AI agents. Beliefs with confidence and decay, governance enforcement, self-healing knowledge, vector search. SQLite-powered, zero-config.

Local-first hierarchical memory brain for AI coding agents — auto-capture, LLM consolidation into beliefs, hybrid retrieval, per-prompt injection, a living Neural Universe monitor, and a daily self-improvement loop. Works with Claude Code, Codex, and any

Tree-based mapping of beliefs, arguments, and evidence. (server link, used by clients)

TypeScript SDK for MnemeBrain — Give your AI a real brain. Agents today store text. Mnemebrain stores beliefs — with evidence, confidence, provenance, and revision logic.

Production AI agents for Toolpack SDK — 8 channel integrations (Slack, Discord, Telegram, SMS, Email, Webhook, Scheduled, MCP), AgentMind persistent cognitive layer (goals, beliefs, reflections), interceptors, evals, and multi-agent coordination

Cortex persistent memory engine for OpenClaw — 4-tier architecture with Bayesian beliefs and people graph

Lifecycle gates for claims and beliefs — four orthogonal axes (truth, retrieval, security, freshness) with per-axis transition rules. Part of Lodestar, the trust layer for AI agents.

CSL style for Journal of Beliefs & Values

CAVE rules engine (spec §24) — forward-chaining `premises => conclusion` rules over current beliefs, with BECAUSE/VIA derivation lineage, noisy-AND confidence, tx-watermark incrementality and idempotent re-runs.

A framework that shares your beliefs

OpenClaw plugin for value mirroring — agents that represent and debate a person's values, beliefs, and worldview.

Contrail ↔ Hermes Agent adapter — recall beliefs and build instruction blocks

Bayesian inference engine. Update beliefs with new evidence. Prior + data = posterior.

This is a component library to be used across the organization that will allow for increased velocity, improve communication and visibility, and a cohesive, over-arching design system for ACE.

Core interfaces of [JaKtA](https://github.com/jakta-bdi/jakta), a Kotlin Multiplatform framework for BDI (Belief-Desire-Intention) agent-oriented programming.

Prolog-based concrete incarnation of [JaKtA](https://github.com/jakta-bdi/jakta), a Kotlin Multiplatform framework for BDI (Belief-Desire-Intention) agent-oriented programming.

Typed facts that outlive the session. Agents assert, recall, and retract through JSON or MCP.

MemOS-oriented Harness scaffold for AI coding agents: AGENTS.md / CLAUDE.md / CODEX.md, Cursor rules, design docs (core-beliefs, ADR, delivery SOP), dependency-cruiser, Husky, freshness & CI scripts. Non-interactive npm CLI copies templates and prints an