AI Ecosystem Directory
A searchable, curated directory of 120+ frontier frameworks, coding agents, model APIs, local runtimes, evaluation suites, and research papers.
AI Coding Assistants & Engineering Agents
Autonomous coding agents, AI-native IDEs, CLI assistants, and repository context tools.
Agent-first IDE from Google where autonomous agents plan, execute, and verify software tasks with artifacts.
Anthropic's terminal-based coding agent for navigating, editing, and executing tasks across large codebases.
AI-native code editor designed for codebase chat, multi-file edits, and agentic refactoring.
AI-first development environment combining code editing with cascade agentic assistance.
High-performance code editor built in Rust with native AI editing and inline agent workflows.
Agentic IDE built around formal spec-driven development and autonomous implementation.
Open-source IDE assistant bringing tab-autocomplete and custom context to VS Code and JetBrains.
Command-line pair programmer that edits files directly in your local git repository.
Open-source command-line agent built around Google's Gemini models for autonomous terminal tasks.
Open-source local coding agent by Block that connects to any model via MCP.
Research agent from the SWE-bench team that autonomously resolves real GitHub issues.
Open-source terminal coding agent with 75+ provider backends, LSP support, and MCP integration.
Autonomous software engineering agent operating inside isolated Docker containers.
Autonomous VS Code extension with command execution, file editing, and browser testing.
Local terminal interface for executing code and automating computer tasks with natural language.
Generates entire software projects from high-level specifications and clarifying questions.
Terminal-based AI coding workflow designed for large multi-file implementation plans.
Agentic coding tool from Sourcegraph built for editor and terminal engineering workflows.
AI assistant for JupyterLab enabling conversational coding in Python notebooks.
AI pair programmer integrated into IDEs, terminals, and GitHub pull requests.
AWS-focused coding assistant for IDEs, terminals, and cloud infrastructure tasks.
AI pull request review assistant for summarizing changes and surfacing bugs.
Open-source automation for PR review, descriptions, code suggestions, and security feedback.
Generative UI system for creating accessible React and Tailwind components from text prompts.
Browser-based AI web development agent for building and deploying full-stack web applications.
Conversational app generation platform turning product specifications into production-ready web apps.
Turns any GitHub repository into an LLM-friendly text digest for rapid context ingestion.
Packs entire source repositories into structured, token-efficient context files for AI models.
Agent Orchestration & LLM Frameworks
Multi-agent graph systems, typed Python libraries, agent protocols, and visual app builders.
Stateful graph framework for building multi-step, multi-agent LLM loops with human approval gates.
Microsoft's framework for building event-driven multi-agent conversations and collaboration networks.
Agent orchestration framework structured around roles, delegable tasks, and sequential processes.
OpenAI's official Python framework for multi-agent workflows with handoffs, guardrails, and tracing.
Production-ready typed Python framework for building structured, validated LLM applications.
Minimalist Hugging Face library for lightweight agents that write and execute Python code.
Microsoft's SDK for integrating LLMs, plugins, and multi-agent workflows into enterprise apps.
Multi-agent framework that converts user requirements into PRDs, designs, and code files.
TypeScript framework for building agents, multi-step workflows, persistent memory, and tools.
Self-improving personal agent platform by Nous Research with memory loops, 20+ platforms, and custom skills.
Open standard for connecting AI models to local tools, databases, and enterprise services.
Open-source library enabling LLM agents to control real browser sessions for web automation.
Browser automation infrastructure built for AI agents with session management and proxy support.
Secure cloud environments for running AI-generated code, shell scripts, and browser tools.
Universal memory layer enabling agents to retain personalized user state across sessions.
Platform for stateful agents with long-term memory management based on MemGPT research.
Comprehensive ecosystem for LLM chains, agent tools, document loading, and production integrations.
Data framework for indexing, retrieving, and connecting LLMs to private enterprise data.
Modular open-source framework for building production search and question-answering pipelines.
Declarative framework for programming LLM pipelines and automatically compiling optimal prompts.
Unified SDK and proxy server to call 100+ LLM APIs using the standard OpenAI client format.
Python library for extracting structured Pydantic outputs from LLM responses with validation.
Guided generation library constraining model sampling to exact JSON schemas, regex, and grammars.
Open-source visual platform for building, testing, and operating LLM applications and workflows.
Open-source drag-and-drop builder for constructing LLM chains, agents, and RAG pipelines.
Visual environment for rapid prototyping of multi-agent and RAG applications.
Document conversion and ingestion toolkit for parsing PDFs, DOCX, and slides into LLM-ready markdown.
API that turns entire websites into clean, LLM-ready markdown or structured JSON datasets.
Open-source asynchronous web crawler tailored specifically for LLM data pipelines.
Local LLMs, RAG & Vector Databases
Local GGUF runtimes, GPU serving engines, local desktop apps, and vector similarity search.
Simple local CLI and background daemon for running open-weight LLMs like Llama 3, Mistral, and DeepSeek.
Cross-platform desktop application for discovering, downloading, and chatting with local GGUF models.
Feature-rich, self-hosted chat interface for local Ollama runtimes and OpenAI-compatible endpoints.
Open-source offline-first desktop AI app that runs models locally on hardware or connects to remote APIs.
Desktop interface for managing, comparing, and running local and cloud language models.
Simon Willison's CLI utility and Python library for prompting local and remote models from the terminal.
High-performance C/C++ inference engine for local quantization and open-weights execution.
Fast, easy-to-use LLM serving engine built around PagedAttention for maximum GPU throughput.
Fast serving framework for execution of LLMs and vision-language models.
Mozilla project packaging a model and llama.cpp runtime into a single cross-platform executable.
Self-hosted OpenAI-compatible REST API backend for running models on consumer hardware.
Apple's array framework for efficient machine learning and LLM inference on Apple silicon.
High-performance C/C++ port of OpenAI's Whisper for on-device speech recognition.
Meta's C++/Python library for efficient similarity search and dense vector clustering.
High-performance open-source vector database and search engine written in Rust.
Open-source vector database supporting hybrid search and modular ML model integrations.
Developer-friendly open-source embedding database for building Python and JS LLM apps.
Embedded vector database for multimodal AI built on the Lance columnar storage format.
Model Providers & API Gateways
Frontier API consoles, open-weights releases, model hubs, and high-speed LPU inference providers.
APIs for GPT-4o, o1/o3 reasoning models, DALL-E, Whisper, and real-time audio.
API console for Claude 3.5 Sonnet, Haiku, and Opus models.
Web console and API suite for Gemini 1.5 Pro, Flash, and multimodal models.
Frontier open-weights reasoning models (DeepSeek-V3, DeepSeek-R1) with sparse MoE architecture.
Meta's open-weight foundation model series available for self-hosting and fine-tuning.
Efficient open-weights and enterprise models including Mistral Large, Codestral, and Pixtral.
Unified gateway providing pay-per-token API access to over 200 open and proprietary LLMs.
Ultra-fast cloud LLM inference engine powered by specialized Language Processing Units (LPUs).
Alibaba's open-weight multilingual model family with strong coding and math benchmarks.
The open platform for sharing machine learning models, datasets, spaces, and tooling.
Evals, Observability, Books & Papers
Tracing platforms, LLM-as-judge test runners, production textbooks, and research papers.
Open-source LLM engineering suite for tracing, prompt management, metrics, and automated evals.
Platform for debugging, testing, evaluating, and monitoring LLM applications.
CLI tool and library for evaluating LLM output quality, red-teaming, and regression testing.
Open-source LLM evaluation framework for writing unit tests on RAG and agent outputs.
Evaluation framework specifically designed for measuring retrieval and generation accuracy in RAG.
Open-source AI observability platform for tracing, evaluations, and vector embedding analysis.
OpenTelemetry-native auto-instrumentation and tracing library for generative AI applications.
Observability, logging, caching, and rate-limiting proxy layer for LLM applications.
Open-source tracing, evaluation, and monitoring platform for LLM production systems.
Standard evaluation benchmark testing language agents on resolving real-world GitHub issues.
Benchmark evaluating terminal-based AI agents across real shell tasks.
Independent benchmarks measuring LLM quality, latency, throughput, and pricing.
Crowdsourced open LLM evaluation leaderboard based on human side-by-side preference votes.
The landmark 2017 Google paper introducing the Transformer architecture.
The foundational research paper defining interleaved reasoning and tool execution in LLM agents.
Research paper detailing reinforcement learning incentives for reasoning models.
Kaplan et al. paper showing how language model performance scales with compute and parameters.
Parameter-efficient fine-tuning paper that became the default for adapting LLMs.
Chip Huyen's comprehensive guide to designing, building, and scaling production AI systems.
Book covering data pipelines, production ML infrastructure, model monitoring, and iterative feedback.
Sebastian Raschka's Manning book building transformer LLMs layer-by-layer in PyTorch.
Manning book detailing agent loops, tool integration, planning, and multi-agent systems.
Andrej Karpathy's legendary video series building micrograd, GPT, and tokenizers from scratch.
Leading podcast and essay series focused on the practical AI engineering stack.