AI Agent Frameworks, MCP Servers and SDKs

ALTK Evolve gives AI agents persistent memory of earlier tasks so they can reuse lessons across sessions. It's for coding assistant users and agent builders who want guidance drawn from past successes and mistakes.

LEANN is a local vector database for people building AI search over personal data or knowledge collections. It reduces index storage by computing embeddings on demand rather than keeping every embedding on disk.

PandasAI is a Python library that executes data analysis code in your environment and lets analysts ask questions in plain language. It can isolate that execution in Docker. Its documented LiteLLM integration uses OpenAI's gpt-4.1-mini, so model requests go to a cloud service and require an API key.

ast-grep MCP lets AI coding assistants find code by syntax structure. It's for developers who need searches for specific constructs or relationships that text matching alone can't express.

OASIS is a social media simulator for researchers studying how information spreads, groups polarize, and users follow the crowd. It uses LLM agents as simulated participants in Twitter and Reddit environments.

Rig is a Rust library for developers building AI agents and LLM applications. It provides a shared interface across model providers and vector stores for applications that need to work with different backends.

Forge Orchestrator coordinates Claude Code, Codex CLI and Gemini CLI working on the same repository, so they can share decisions and avoid conflicting edits. It's built for internal use at NXTG; outside use isn't supported, there's no SLA, and interfaces may change without notice.

Rivet AgentOS is a runtime for developers building applications where AI agents need a working environment for code and tools. It runs inside an existing backend process, with WebAssembly isolation instead of a separate virtual machine for each agent.

OpenContracts is a self-hosted document intelligence platform for teams working with legal archives, research libraries or internal policies. It connects documents through citations and annotations, giving people and AI agents a shared record of how source material relates. The code is open source under the MIT license; the public website also provides collections readers can browse.

Memorix keeps project memory for AI coding agents across chats and switches between tools. It's for developers who want decisions, fixes and unfinished work to remain available to the next session.

txtai is a Python framework for developers building search, chat with their data, and AI agents on their own hardware. It combines a searchable knowledge store with model workflows, so applications can retrieve context and act on it.

LlamaIndex is a Python framework for developers building AI agents and applications that answer questions using their own data. It provides the connections and retrieval tools needed to give a language model relevant context from documents and other sources.

Pi Subagents is an extension for developers who want to divide coding work among specialized AI agents in pi. Agents run in separate sessions, so you can assign distinct roles and follow their progress individually.

Jsonformer is a Python library for developers who need structured JSON from language models running through Hugging Face Transformers. It controls the JSON structure while the model supplies the values, reducing reliance on prompts that ask a model to produce correctly formatted output.

Laminar helps developers inspect AI agent runs, find recurring failures and test fixes. It connects production traces with debugging and evaluation tools.

LLMLingua is a Python library for developers building LLM applications that need to fit lengthy input into a model's context window or reduce token costs. It uses smaller models to remove less useful text before the prompt reaches the answering model.

MASFactory is a Python framework for developers building workflows in which several AI agents collaborate. It combines natural-language design, code authoring and visual editing so teams can define the work and inspect how agents carry it out.

OpenRath is a Python runtime for developers building multi-agent workflows that need an inspectable record of their work. It keeps context and evidence together so engineering and research teams can trace how a result was produced.

OpenShell is a self-hosted runtime for teams that need autonomous AI agents to work with files and APIs under explicit access limits. It runs agents in isolated sandboxes and enforces policies at the kernel level.

llama.rn is a library for developers adding local LLM features to React Native apps on iOS and Android. It brings llama.cpp inference onto the device, so an app can process model requests on the user's hardware.

OpenRig coordinates coding agents for developers working on projects that last weeks or months. You talk to a lead agent, which delegates to specialists and brings back results or decisions that need your attention.

FastRTC is a Python library for developers building voice assistants, webcam AI apps, and other interactive audio or video tools. It connects live media streams to Python functions so developers can focus on how their application responds.

mcp-proxy connects AI clients and Model Context Protocol (MCP) servers that use different connection types. It's for developers who need desktop clients to reach remote tools, or want other clients to access a local MCP server.

MEX keeps project knowledge alongside code so engineers and their AI coding agents can share context across teammates and sessions. Architecture, requirements and decisions live in readable Markdown that travels with the repository through Git.

LeanCTX helps developers control the information their AI coding tools receive. It runs locally between those tools and project files or command output, while your existing application keeps control of the model and workflow.

All-in-One Sandbox gives AI agents a place to browse websites, execute code and work with files. It's for developers building agents that need these tools together, with a shared filesystem instead of separate environments for each task.

Velocut is a local video editor and 3D scene studio for creators who want to edit manually or delegate work to Codex. Both work on the same project, with a shared undo history.

LoopX is an open-source control plane for AI agents working on multi-day coding, research and personal projects. It preserves the goal and decisions between sessions so work can resume after a pause or handoff.

mcpo is a self-hosted proxy for developers connecting MCP tools to AI agents and apps that accept OpenAPI servers. It exposes existing MCP servers through HTTP, avoiding a separate adapter for each tool.

SymForge is a local MCP server for developers using AI coding agents on large repositories or long source files. It gives agents focused views of symbols and their relationships so they can investigate code without reading entire files.

Haystack is a Python framework for developers building AI agents, document search and applications that answer questions using their own data. Its pipelines give teams control over the information a model receives and the steps that produce an answer.

IntentKit is a self-hosted AI agent framework for developers who need several agents to share work and carry out recurring tasks. It runs as a server application, with chat channels and APIs for interacting with the agents.

Sandcastle is a TypeScript library for developers building automated coding workflows in scripts, CI pipelines, or custom tools. It coordinates sandboxed AI coding agents and manages where their Git changes land.

PAL MCP Server connects coding assistants to other AI models for second opinions, code review and planning. It's for developers who want to consult several models within the same coding session, using clients such as Claude Code, Codex CLI, Gemini CLI or Cursor.