Open Source

Harness Router is a decision layer for AI coding agents that makes tool selection faster, cheaper, and more reliable before execution. Use it as a Codex hook to automatically route every tool call, or as a skill when the agent should invoke routing explicitly. Obvious decisions stay on the fast path, ambiguous choices are resolved with Jev, and complex multi-step decisions can escalate to MCTS.

Your Markdown lives in a dozen folders. Most of the time you don't want to write — you just want to read one section. Every Markdown tool is a writing app with a preview pane. MuM is the opposite: a native macOS reader with no web engine inside — all AppKit. Several projects open at once, each remembering where you stopped. Cross-project search. Typesetting tuned for CJK as well as Latin. ~0.3s cold start, 1.7 MB download, 100+ fps scrolling a 5 MB document. Open source, MIT.

fleeting.chat is an HTTP channel for agent-to-agent messaging (2–8 ED25519 seats). Humans hit Generate on the site; agents follow https://fleeting.chat/llms.txt — reserve/join, send, poll (optional long-poll), optional base64 files. No accounts for basic use. Channels expire by default (idle + absolute TTL; defaults 24h idle / 48h absolute, configurable up to 30d). Optional server at-rest encryption (AES-GCM; not E2E). Open source (ISC).

Szept is a light, fast Mac app for talking to AI agents: dictate in 60+ languages, clean up text with AI, capture and record the screen, and hand it to Claude Code. Source code included.

Build and test conversational workflows with Jev. See why each step happened, save conversations as tests, and catch wrong turns before you ship. Export runnable TypeScript, workflow JSON, and an integration skill for your coding agent. Free and open source. Bring your own key for live runs.

Prowler brings open-source cloud security into the agentic era. It combines deterministic, auditable security checks with AI-powered investigation, prioritization, reporting, and remediation. Prowler supports AWS, Azure, GCP, Kubernetes, Microsoft 365, GitHub, IaC and more, with thousands of checks, compliance mappings and attack paths, plus Lighthouse AI, MCP support and advanced compliance workflows. Connect your agent now to start defending your cloud at cloud.prowler.com.

QuietGlass is a free, open-source Mac app for screen privacy. Look away with compatible AirPods and your screen blurs. Your camera can spot another face in view and warn you or blur the screen, even while you're still looking. Protect a window or selected area, blur sensitive text, or keep your active window clear with Focus mode. Optional owner recognition checks it's you before clearing camera protection. Native Swift, notch controls, and processing on your Mac. macOS 14+.

Fast, free, drop-in multilingual profanity and toxicity screener for Node.js, powered by System 1 models like TypeSafe AI Jev and Laya. Catches leetspeak, ASCII drawings, character spacing, and romanized profanity across all languages including Kannada, Telugu, Tamil, Hindi, and Bengali. Ultra-low cost • Multilingual • Native Indic support • Evasion-aware • Sub-500ms • ~$0.000004/message • Configurable moderation actions • Open source (npm i gg-friggin-ez)

Fulvid is an open source Markdown and MDX editor for open knowledge. Keep your knowledge in portable Markdown files on disk, ready to read, link, reuse, version, and work with across tools and LLM workflows. Search, preview, document links, and Graph navigation stay close to the filesystem. No proprietary vault or cloud workspace.

SereneDB is the result of 12 years of development - an open-source database that does ultra-fast full-text and fast analytics in one engine. Postgres- and Elastic-compatible: keep your SQL, your drivers, and your Elastic clients; drop the second system and the ETL between them. In our public benchmark, it outperforms Elasticsearch, ClickHouse, and Postgres search extensions, and indexes 1B logs in under 8 minutes at ~10x less disk. Apache 2.0, methodology, and raw results are public.

Taiafox filters a hundred incoming messages down to the three that matter, coordinates responders by voice, and re-plans in under a second when the fire turns.

KnownPath is a shared knowledge network for AI coding agents. When an agent solves a reusable technical problem, it can contribute the generalized fix so other agents can find it later instead of debugging from scratch. KnownPath adds trust scoring, privacy-safe contributions, deduplication, and real outcome feedback so useful fixes get stronger over time. It works through MCP + an Agent Skill across Codex, Claude Code, Cursor, Gemini CLI, and OpenCode.

Antifailure builds a throwaway copy of your production stack for each branch: the real schema and data volume with every identifier masked and the masking proved, a network the app cannot escape, and agents that use it like people. Your change runs there first, it tells you what would have broken, and then the copy deletes itself. MIT licensed, one command, no account.

Matrix OS gives you and your AI agents a persistent computer in the cloud. Run Claude Code, Codex and other agents without tying them to your laptop. Keep your files, repos, terminals and apps running, close your laptop, and come back from any device. Run multiple agents in parallel, or self-host Matrix for full control.

Build with our CLI and dev tools. Test on simulators made for agents. Ship with the best CI/CD for mobile and learn from every release.

Per-project semantic search over your Claude Code session history — not a shared or centralized memory store. Indexes JSONL transcripts into a local SQLite file, retrieves via bge-m3 embeddings (Ollama) with automatic FTS5 BM25 fallback, and auto-injects relevant past context at session start. Companion to session-end: session-end gives "where I left off," session-indexer gives "what we decided" — same Stop/SessionStart hooks. Apache 2.0, 75 passing tests, go install and nothing to deploy.

Stop rebuilding wearables from scratch. SensWear is an open-source, modular platform for creating smart rings, wristbands, patches, and custom devices. Mix sensor modules, access synchronized raw data, run edge AI, and build with Python or TypeScript without vendor lock-in or a required cloud. The full stack is MIT-licensed. Back the ready-to-use development kit on Crowd Supply: https://www.crowdsupply.com/sensera-technologies/senswear-wearable-dev-kit

Warren runs agent harnesses as isolated, observable workloads on infrastructure you control. It owns the workspace, limits, live events, recovery, and Git delivery.

Evidence is a code-based business intelligence tool. It's the best way to build live dashboards and reports with coding agents.

Managing markdown based context comes with challenges like context poisoning, contradictions, non-determinism and context navigation difficulties. Over time, we built out a structured context management layer at our company (DrDroid) - with Open Index, we are sharing it with the ecosystem!