Model gateways, inference, agent frameworks, memory, evaluation and observability for LLM applications.
Autonomous data infrastructure that turns strategic objectives into verified datasets and live intelligence streams.
OpenAI visual canvas for composing workflows of agents, tools and control-flow nodes, deployable through ChatKit or the Agents SDK.
Agent-MCP is a framework for creating multi-agent systems that enables coordinated, efficient AI collaboration through the Model Context Protocol (MCP).
TypeScript framework from Inngest for building single agents through multi-agent systems with tools, orchestration at its core, and MCP servers usable as tools.
Developer platform for tracing, debugging and monitoring agents with event tracking, time-travel replay, audit trails and cost tracking across many frameworks.
Python SDK and service for building agents that automate browser UI workflows with human-in-the-loop approval, and integrate tools via API calls or remote MCP.
Agent platform for building agents that run on schedules or triggers to handle repeatable workflows such as research, content drafting, monitoring and triage.
Anthropic SDK for building agents on the Claude Code harness with tools and MCP servers.
Open-source AI memory platform connecting scattered sources into a knowledge base so agents recall prior work and cite answers, with an MCP server for read/write.
Model Context Protocol (MCP) server for Opik, the open-source LLM observability and evaluation platform, built by Comet.
Framework and platform for multi-agent crews and flows with MCP server integration.
Open-source LLM evaluation framework with agentic metrics for task completion, tool correctness, plan adherence and trajectories; integrates with Confident AI's MCP server.
JVM framework for agentic applications that blends LLM calls with code and domain models, plans dynamically with GOAP, consumes MCP servers and can act as an MCP server.
Google Cloud managed service to deploy, operate and scale agents with a serverless runtime, sessions and memory, evaluation, code sandboxes and MCP connectivity.
Embeds specialized agents for planning, code review, security analysis and CI/CD across GitLab workflows, with MCP client support for compatible tools.
Enterprise AI platform connecting company data sources for search and knowledge discovery, letting teams build agents that act on enterprise knowledge.
Python framework plus visual Nodes and Cloud deployment for building agents, multi-agent systems, pipelines, workflows and RAG with chat memory.
AI reliability framework and platform with runtime guardrails that detect policy violations, hallucinations and data leakage, plus generated evaluation datasets.
Open-source LLM observability platform to route, debug and analyze applications with request tracking, sessions, prompt management, rate limits and alerts.
Agent observability and evaluation platform tracing prompts, tool calls and decisions, with online evals and drift alerting, plus a HoneyHive MCP and CLI.
Hugging Face MCP server for the Hub, models, datasets, papers and Gradio applications.
Gateway governing LLM, MCP and agent-to-agent traffic with MCP server generation, authentication controls and A2A metrics.
Open-source agent observability platform capturing LLM and tool calls, surfacing failure modes as Signals, and building eval datasets; includes a CLI and MCP.
Open-source framework for building agents with any model provider using proven agent patterns while keeping control over harness, context, memory and model choice.
Visual framework for building agents and RAG applications that can run as MCP servers or clients.
Agent observability and evaluation platform providing tracing, monitoring and performance scoring for visibility into agent behavior.
Open-source AI gateway unifying many LLM providers behind one API with spend and access controls, and routing agents and MCP servers through the same gateway.
Realtime framework for voice, video and physical AI agents in Python and Node.js that join LiveKit rooms with tool use and multi-agent handoffs.
TypeScript framework for agents and workflows, with MCP clients for external tools and servers for exposing Mastra capabilities.
Persistent memory layer for agents that extracts, compresses and retrieves memories across sessions to cut tokens and latency, with an MCP integration.
Multi-agent framework that assigns product manager, architect and engineer roles to LLM agents that follow SOPs to turn a one-line requirement into software deliverables.
Microsoft framework for building multi-agent applications in .NET and Python with tool and MCP integrations.
Open-source toolkit for programmable guardrails on LLM applications with input, dialog, retrieval, execution and output rails.
OpenAI framework for multi-agent workflows with handoffs, guardrails and MCP servers.
Unified API to hundreds of models with routing and fallbacks, plus an Agent SDK for multi-turn agents with tool execution and MCP tool support.
Open-source framework for voice and multimodal agents with real-time orchestration, interruption handling, multi-agent workflows and WebRTC, SIP and PSTN transports.
AI gateway and observability platform with an MCP gateway for agent tools.
Python agent framework from the Pydantic team with typed tools and MCP client support.
Agent framework built on Qwen models featuring function calling, MCP tool integration, code interpreter, RAG and example browser-assistant applications.
Rust library for LLM apps and agents with one API across many providers, type-safe tools, streaming, RAG pipelines, multi-agent workflows and MCP server integration.
Persistent memory for AI agents — log and recall conversation context over MCP.
Enterprise agent platform with governed RAG over multimodal documents, hallucination detection and policy enforcement, deployable in SaaS, VPC or on-premise.
Enterprise AI platform and control plane for models, data, guardrails and observability, with Agent Builder, SDKs and data connections via connectors and MCP.