#llm

Local-first AI workflows for engineers.

Local-first AI workflows for engineers.

Detent is board-driven agentic work orchestration, shipped as a single Go binary — it watches a GitHub Projects board and runs each ready issue through an isolated Git worktree, a coding agent, your validation gate, and a serialized merge train, live on web and terminal dashboards.

Web dashboard for Hermes Agent — multi-platform AI chat, session management, scheduled jobs, usage analytics

A WebRTC-native, audio-first conversational-AI framework for Go.

💬 A performant virtual chat component for Svelte applications - efficiently render large scrollable chat interfaces with minimal memory usage

An Agent Development Kit (ADK) allowing for seamless creation of A2A-compatible agents written in Go.

vMLX - JANGTQ Uber Compressed MLX Models - L2 Disk Cache (survives restart) + L1 Paged (super fast ttft) + Hybrid SSM Scheduler + Cont Batching + etc!

PostgreSQL compiler. One source, many faithful forms.

Research-grade investment decision engine for AI agents: isolated multi-agent committee, auditable verdicts, backtests with lookahead protection, published negative results

LLM prompt language based on Jinja. Banks provides tools and functions to build prompts text and chat messages from generic blueprints. It allows attaching metadata to prompts to ease their management, and versioning is first-class citizen. Banks provides ways to store prompts on disk along with their metadata.

A universal git-native AI agent framework. Your agent lives inside a git repo — identity, rules, memory, tools, and skills are all version-controlled files.

Context engineering for AI agents. ~80% fewer tokens. Fix tool overload. Skills and memory with in-process BM25 and semantic retrieval. Progressive Disclosure. No vector DB.

Google Gemini platform bridge for Symfony AI

Ollama platform bridge for Symfony AI

OpenRouter platform bridge for Symfony AI

Symfony bridge for AI Mate - provides Symfony container introspection and optional profiler data access tools

AI partners for every person: multi-user, multi-agent memory, trusted tools, schedules, sandboxed workspaces, and chat.

TurboQuant+ KV cache compression for vLLM. 3.8x smaller KV cache, same conversation quality. Fused CUDA kernels with automatic PyTorch fallback.

LiteLLM for Go, the easiest way to write LLM-based programs in Go

Agentic terminal AI assistant with a Bubble Tea TUI, code-graph review, semantic search, multi-provider LLMs, and MCP tools.

Cognitive memory engine for AI agents — temporal decay, contradiction detection, autonomous consolidation, knowledge graph, ANN recall via HNSW. Embeddable Rust library with Python bindings; powers yantrikdb-server (HTTP gateway, MCP server, openraft cluster). AGPL.

Production-grade Go SDK for building AI agents with long-term memory, knowledge retrieval, and voice — runnable as a library, a daemon, or a real-time pipeline.

A Javascript SDK for the Linkup API

Verifies & compiles reduced-motion-safe, on-budget UI animation for AI-generated web apps. Validated specs → deterministic GSAP/WAAPI-CSS; WCAG 2.2.2 + 2.3.3 checks. MIT core + keyless MCP server. Not Motion.dev/Framer Motion, not the Android/iOS MotionSpec classes, not a video generator.

Runtime Firewall for AI agents which catches the rogue tool call before it runs. Dangerous commands, secret leaks, prompt injection. For Claude Code, Codex and framework SDKs

Attention is your scarcest resource. Chief is the local-first layer that guards it — turning every agent, alert, and feed into one honest call: interrupt, or not.

🐙 Drop-in tools that give AI agents reliable, permission-aware access to external systems.

OpenTelemetry instrumentation plugins for JavaScript-based AI coding agents. Collect traces, tool calls, and LLM metrics from Claude Code and OpenClaw — zero code changes required.

Cheap DeepSeek coworker for Claude Code/Codex. Delegate bulk reads, boilerplate, and chat extraction to save 60-70% of your token budget.