Python

Flue is an open-source Python tool that lets coding agents control desktop software from the command line without MCP servers. The README says it works with apps like Photoshop, Blender, Unity, Houdini, and Microsoft Office, and supports harnesses such as Claude Code, Codex, Gemini CLI, Cursor, and OpenCode.

A CROWDB blog post showing how to run an Iceberg table stack in one Docker container, write sample data with PyIceberg, and inspect the results in pandas and the local console.

Open-source Python safety layer for AI agent actions. It checks proposed actions before execution, supports approvals and blocking rules, and includes a demo, receipts, and docs for running locally or with Postgres.

A local zero-dependency sandbox for agents that records files, commands, database activity, network access, and screenshots for review before approving changes.

Dependency-free Python library that compiles a declarative JSON query format into safe, parameterized SQL for LLM workflows, REST APIs, and frontend-backed data tables.

A local CLI tool that keeps Claude Code sessions warm by refreshing prompt cache during long subagent runs, aiming to reduce re-encoding costs in extended sessions.

A Python framework for local document search that replaces traditional embedding-based RAG with BM25 indexing, Meilisearch, and an agent loop for search, fetch, and citation-backed answers.

An open-source Python tool that turns multi-persona AI debate into pre-registered, time-anchored predictions and scores each seat against reality with Brier-style calibration. It includes a CLI, sample workflows, docs, and examples, plus a BYOK design for running on your own model key.

A Python/Flask code example that shows how to build a retrieval-augmented generation API with Telnyx embeddings and chat completions. The repo includes setup steps, environment variables, endpoint docs, and sample responses.

Open-source persistent memory agent for AI coding tools and assistants. It supports local and cloud setups, integrates with several agent IDEs and CLIs, and offers remember, recall, and answer primitives for long-term context.

MCP server for Unreal Editor that exposes a small MCP surface for a large set of editor automation tools. It supports blueprint editing, asset management, animation authoring, trace analysis, PIE actions, and Python execution, with setup notes for Unreal Engine 5.x and MCP-compatible clients.

Modular framework for building trading systems with a C++ core, polyglot strategy bindings, and an MCP server for AI coding agents.

Lightweight AI inference engine written in Rust for embedding local LLMs in mobile and desktop apps, with bindings for Python, Flutter, and React Native.

Research code for improving agent harnesses through regularized recursive self-improvement, with evaluation, selection, pruning, and domain-specific adapters.

A step-by-step tutorial that walks readers through making a first Anthropic Claude API call in Python, including environment setup, client creation, token usage, and printing the response.

Open-source LLM uncertainty and hallucination detection tool that estimates epistemic risk in microseconds using exact-match normalized entropy, with Rust gateway and Python package support.

Ridge is a resource mesh for AI agents, with MCP, CLI, and Python access to local projects, Docker, SSH machines, and S3 storage. It focuses on delegated resource access, job coordination, and keeping artifacts out of model context.

Myli is a provider-neutral Python harness for multi-agent visual design workflows. It defines typed design specs, validates JSON candidates at runtime, supports tools and middleware, and can render candidates for visual review with event and trace reporting.

AetherGrid is a Python-based distributed compute orchestration project with API key auth, job lifecycle management, reconciliation, leases, and ADR-driven architecture notes. The snapshot shows a substantial repository with documentation, tests, Docker support, and detailed runtime and security decisions.

A ProtoLink example that stages autonomous agents in a fictional liability tribunal to study how agent-to-agent communication changes decisions, with replayable traces and HTML reports.

Open-source Python library that lets developers integrate an LLM into the Python runtime so missing attributes, methods, and typed values can be inferred on demand with confidence scores.

Banksia is an open-source Python project for building and running accountable AI teams. The repo shows a CLI install flow, a local console, workflow studio, run monitoring, and support for reusable teams and durable state.

OpenSmith is a local-first, open-source alternative to LangSmith for tracing Python LLM pipelines. It supports decorators, context managers, console output, autopatching, CLI commands, local SQLite storage, and optional Postgres and OpenTelemetry integrations.

Python toolkit for building LLM agents with MCP servers, skills folders, memory, streaming, and metrics support.