All tools

Anthropic's most capable widely released model, built for demanding reasoning and long-horizon agentic work.

Claude Haiku 4.5

Fast, cheap small model for classification, routing, high-volume work.

Claude Opus 4.7

Anthropic's deepest-reasoning model; planning, complex synthesis.

Claude Opus 4.8

Anthropic's Opus-tier model (May 2026) with improved coding, agentic, and honesty behavior over Opus 4.7.

Claude Sonnet 4.6

Anthropic's workhorse coding/agent model; strong tool use, long context.

Cohere Command R+

Cohere's RAG-optimized model with strong tool use + citation grounding.

Cohere Rerank v4.0

Cohere's current-generation reranking models (Pro and Fast variants) for boosting retrieval relevance in RAG and agent search.

DeepSeek V3

Open-weights model with strong reasoning at very low inference cost.

DeepSeek V4

DeepSeek's V4 generation of open-weight MoE models with 1M-token context and agent-focused reasoning modes.

Gemini 2.5 Flash

Google's fast, cheap small model; long context, native multimodal.

Gemini 2.5 Pro

Google's frontier multimodal model with very long context.

Gemini 3.5 Flash

Google's stable frontier model for sustained agentic and coding work, the current GA workhorse of the Gemini 3 family.

GLM-5

Zhipu AI's MIT-licensed 744B-parameter MoE model targeting complex systems engineering and agentic tasks.

GPT-5

OpenAI's frontier model; broad capability, strong tool use.

GPT-5.4

OpenAI's more affordable frontier model for coding and professional work, sitting below GPT-5.5.

GPT-5.5

OpenAI's flagship frontier model, positioned as a new class of intelligence for coding and professional work.

Grok 3

xAI's frontier model with real-time X data access.

Grok 4.3

xAI's flagship API model, billed as its most intelligent and fastest, with strong agentic tool calling.

Kimi K2.6

Moonshot AI's flagship 1T-parameter open-weight MoE model with native multimodal input and interleaved-thinking tool use.

Llama 4 Maverick

Meta's open-weights frontier model; self-hostable on any infra.

Mistral Large

Mistral's frontier model — strong multilingual + EU-data-residency story.

Qwen3.5 (397B-A17B)

Alibaba's Apache-2.0 open-weight Qwen3.5 flagship MoE with very low active-parameter cost and broad multilingual coverage.

Qwen 3 Max

Alibaba's open-weights frontier model; very strong on Chinese + code.

OpenAI text-embedding-3-large

OpenAI's most capable embedding model for English and non-English retrieval, served via usage-priced API.

Voyage-3

General-purpose text embedding model tuned for retrieval quality.

Voyage-3-large

Higher-capacity Voyage embedding model for retrieval-quality-sensitive RAG.

Agent Framework 16

SDK / library structuring agent loops, tool calls, state.

AG2

Community-maintained multi-agent conversation framework forked from Microsoft's AutoGen.

Agno

Python framework and high-performance runtime for building and running multi-agent systems.

Claude Agent SDK

Anthropic's official SDK for building production agents on Claude.

CrewAI

Role-based multi-agent orchestration framework for building collaborative autonomous agent crews.

DSPy

Open-source Python framework for programming, rather than prompting, language models.

Google ADK

Open-source, code-first toolkit from Google for building, evaluating, and deploying AI agents.

Haystack

Open-source Python orchestration framework from deepset for production LLM and RAG applications.

LangGraph

Graph-based agent orchestration framework from the LangChain team.

LlamaIndex

Data framework for connecting LLMs to private data, with RAG, agents, and event-driven workflows.

Mastra

TypeScript-first agent framework with workflows, RAG, memory, evals.

OpenAI Agents SDK

OpenAI's official SDK with handoffs, tools, guardrails, tracing.

Pydantic AI

Type-safe Python agent framework built by the Pydantic team around validated structured outputs.

Semantic Kernel

Microsoft's open-source SDK for embedding LLM agents and plugins into .NET, Python, and Java apps.

smolagents

Hugging Face's minimalist agent library where agents act by writing and running code.

Strands Agents

Open-source, model-driven SDK from AWS for building AI agents in a few lines of code.

Vercel AI SDK

Free, open-source TypeScript toolkit from Vercel for building AI-powered apps and agents.

Harness / IDE 14

Runtime surface (Claude Code, Cursor, custom backend).

Aider

OSS AI pair programmer for the terminal; multi-model, git-native.

Amp

Frontier coding agent originally built at Sourcegraph, now developed by the independent Amp Frontier Corporation.

Claude Code

Anthropic's official agentic CLI; plugins, skills, hooks, MCP, subagents.

Cline

Open-source coding agent that runs as an IDE extension and CLI with bring-your-own-key model access.

OpenAI Codex CLI

OpenAI's official terminal coding agent; adopts the SKILL.md standard.

Cursor

AI-native code editor (VS Code fork) with agent mode, codebase indexing.

Devin

Cognition's autonomous cloud software engineer that plans, codes, tests, and opens pull requests on its own.

Factory

Agent-native software development platform from Factory AI, powered by its Droid agents in the terminal, desktop, and web.

GitHub Copilot

GitHub's AI pair programmer with an in-IDE agent mode and a cloud coding agent that turns assigned issues into pull requests.

Goose

Open-source, on-machine AI agent created by Block, now hosted at the Linux Foundation's Agentic AI Foundation.

Jules

Google's asynchronous coding agent that takes GitHub tasks, proposes a plan, and delivers pull requests.

OpenCode

Open-source AI coding agent for the terminal, IDE, and desktop, maintained by Anomaly.

OpenHands

Open-source, model-agnostic platform for running autonomous software engineering agents, formerly known as OpenDevin.

Windsurf (now Devin Desktop)

Agentic AI IDE acquired by Cognition in 2025 and rebranded Devin Desktop in June 2026.

Memory 6

Persistent state across turns / sessions / users.

Cognee

Open-source memory platform that gives AI agents persistent long-term memory via combined knowledge graphs and vector embeddings.

Graphiti

Open-source temporal knowledge-graph framework from Zep for building agent memory.

Letta

Stateful agent memory framework (formerly MemGPT); core+archival tiers.

mem0

Per-user long-term memory layer for agents with fact extraction.

Supermemory

Closed-source universal memory API that adds long-term, personalized memory to AI agents and apps.

Zep

Temporal knowledge graph memory for agents (Graphiti under the hood).

Retrieval / Vector Store 11

Embedding store, hybrid search, RAG infrastructure.

Chroma

Open-source embedding database; popular for local dev and small prod.

Elasticsearch

Distributed search engine whose dense_vector kNN and hybrid retrieval support vector and RAG workloads.

LanceDB

Open-source embedded multimodal vector database built on the Lance columnar format.

Marqo

End-to-end vector search engine that generates, stores, and retrieves embeddings for text and images through a single API.

Milvus

Open-source, cloud-native vector database for large-scale ANN search, with Zilliz Cloud as its managed offering.

pgvector

Postgres extension for vector similarity search.

Pinecone

Managed vector database; the incumbent for production RAG.

Qdrant

Rust-based vector similarity search engine and database with payload filtering.

turbopuffer

Serverless vector + full-text search with object-storage backend.

Vespa

Search and AI serving platform combining vector, text, and structured-data retrieval with machine-learned ranking.

Weaviate

Open-source, cloud-native vector database in Go that stores both objects and vectors.

Eval 7

Test harness, golden sets, regression detection.

Braintrust

Eval and observability platform purpose-built for LLM applications.

DeepEval

Open-source LLM evaluation framework with pytest-style testing and 50+ research-backed metrics, maintained by Confident AI.

LangSmith

Eval, tracing, and prompt management from the LangChain team.

Maxim AI

End-to-end platform for simulation, evaluation, and observability of AI agents and applications.

Patronus AI

Automated LLM evaluation and guardrails platform with proprietary evaluation models.

promptfoo

Open-source CLI for prompt and LLM eval, including red-teaming.

Ragas

Open-source Python framework for evaluating RAG pipelines with model-based, largely reference-free metrics.

Observability / Tracing 9

Run traces, prompt logging, cost tracking.

AgentOps

Python SDK and hosted platform for monitoring, debugging, and cost-tracking AI agents.

Arize Phoenix

Open-source LLM tracing and evaluation platform from Arize AI, built on OpenTelemetry.

Galileo

LLM and agent evaluation and observability platform for testing, monitoring, and guardrailing GenAI apps.

Helicone

LLM observability and cost tracking via proxy or async logging.

Langfuse

Open-source LLM observability, eval, prompt mgmt.

Langtrace

Open-source, OpenTelemetry-based end-to-end observability tool for LLM applications.

LangWatch

Open-source LLM observability, evaluation, and AI agent testing platform.

Traceloop

OpenLLMetry — OpenTelemetry-based LLM tracing.

Weights & Biases Weave

LLM tracing and evaluation toolkit from Weights & Biases for developing generative AI applications.

Prompt Management 4

Versioned prompts, A/B testing, prompt registry.

Agenta

Open-source LLMOps platform combining prompt management, evaluation, and observability.

Latitude

Open-source prompt engineering platform with versioning and evals.

Opik

Open-source LLM evaluation and observability platform from Comet with built-in prompt versioning and optimization.

PromptLayer

Prompt registry and logging for production LLM apps.

Sandbox / Code Execution 4

Safe code execution environments for tool calls.

Daytona

Secure, fast-starting sandbox infrastructure for running AI-generated code and agent workloads.

E2B

Cloud sandboxes for running AI-generated code safely.

Modal

Serverless compute for Python with GPU and sandbox primitives.

Runloop

Devbox sandbox infrastructure purpose-built for running and evaluating AI coding agents.

Deployment Target 5

Where the agent runs (serverless, containers, edge).

Cloudflare Workers

Edge serverless platform with AI primitives and Durable Objects.

Fly.io

Container hosting near users; common for stateful agent backends.

Inngest

Durable functions and event-driven queueing platform for agent backends, with the AgentKit framework for building AI agents.

Temporal

Durable-execution runtime that persists workflow state and automatically retries failures, used to run long-running agent workflows reliably.

Vercel Functions

Serverless functions on Vercel; common for Node-first agent backends.

MCP Servers 50

Model Context Protocol servers exposing tools/resources.

Airtable MCP

Airtable MCP — table search, list, row CRUD across bases.

Apify MCP Server

Official Apify MCP server exposing thousands of Apify Actors as web-scraping and automation tools.

Arcade

MCP runtime and authorization platform that lets agents call tools as authenticated end users.

Asana MCP Server

Asana MCP — tasks, projects, comments via the Asana API.

Atlassian Rovo MCP Server

Official Atlassian remote MCP server for Jira, Confluence, and Compass.

AWS MCP Servers

Official suite of AWS MCP servers exposing AWS docs, services, and best practices to agents.

BigQuery MCP

BigQuery MCP — schema discovery, NL2SQL, query execution against GCP warehouses.

Brave Search MCP

Privacy-focused web search via Brave's independent index.

CircleCI MCP Server

CircleCI MCP — build logs, test insights, flaky-test detection, pipeline runs.

ClickHouse MCP Server

Official ClickHouse MCP server for schema discovery and analytics SQL queries against ClickHouse.

Cloudflare MCP Server

Cloudflare MCP — Workers, DNS, R2, KV, D1, analytics via agent calls.

Composio

Tool-integration platform exposing 1,000+ app toolkits to AI agents with managed authentication and MCP support.

Context7 MCP

Inject version-specific docs for 50+ frameworks into agent context.

Datadog MCP Server

Datadog MCP — metrics, logs, traces, incident mgmt via one MCP interface.

Discord MCP Server

Community MCP server for Discord channels, messages, and members.

Exa MCP Server

Agent-focused semantic web search with structured-result returns.

Fetch MCP Server

Reference MCP server for fetching and converting web content for LLM consumption.

Figma MCP Server

Figma MCP — read frames, components, comments; export assets to feed design-to-code agents.

Filesystem MCP Server

Reference MCP server for secure local file operations with configurable access controls.

Firecrawl MCP Server

Web scraping + search + crawl via Firecrawl, returning agent-ready content.

Git MCP Server

Reference MCP server for reading, searching, and manipulating Git repositories locally.

GitHub MCP Server

Official MCP server exposing GitHub repos, issues, PRs, actions.

Gmail MCP Server

Gmail MCP — read, search, draft, send, label messages via the Gmail API.

Google Calendar MCP

Google Calendar MCP — list/create/update events; check availability for scheduling agents.

Google Drive MCP

Google Drive MCP — search, read, and write files in a user's Drive.

HubSpot MCP Server

HubSpot MCP — contacts, companies, deals, sequences for CRM-bound agents.

Intercom MCP Server

Intercom MCP — conversations, contacts, articles for support agents.

Linear MCP Server

Official MCP server for Linear issues, projects, and cycles.

Memory MCP Server

Reference MCP server implementing a knowledge-graph persistent memory.

MongoDB MCP Server

Official MongoDB MCP — query, aggregate, manage collections, Atlas admin.

Notion MCP Server

Official MCP server for Notion pages, databases, and search.

PagerDuty MCP

PagerDuty MCP server — incidents, schedules, services, on-call ack/escalate.

Playwright MCP Server

Microsoft's MCP server exposing Playwright browser automation.

Postgres MCP Server

MCP server for Postgres — schema introspection + read-only SQL by default.

Redis MCP Server

Redis MCP — key/value ops, pub/sub introspection, RedisJSON, RediSearch.

Salesforce MCP Server

Salesforce MCP — Accounts, Opportunities, Cases, custom objects via SOQL/MCP.

Sentry MCP Server

Official Sentry MCP server — error reports, stack traces, releases, 16+ tools.

Sequential Thinking MCP

Reference MCP server for structured multi-step reflective reasoning.

Shopify MCP

Shopify's official MCP servers for store catalog/cart (Storefront) and API/app development (Dev).

Slack MCP Server

Reference MCP server exposing Slack channels and messages.

Smithery

MCP server registry and hosting platform connecting agents to a catalog of 6,000+ tools and integrations.

Snowflake MCP

Snowflake MCP — natural-language queries against the warehouse with role-based access.

Stripe MCP Server

Stripe MCP — customers, payments, subscriptions, invoices via natural-language ops.

Supabase MCP Server

Official Supabase MCP server — Postgres, auth, storage, edge functions, project mgmt.

Tavily MCP Server

Real-time web search + extract optimized for citation-heavy research agents.

Time MCP Server

Reference MCP server for current time and timezone conversion.

Twilio MCP Server

Alpha open-source MCP server from Twilio Labs exposing Twilio's messaging and voice APIs as tools.

Vercel MCP

Vercel MCP — projects, deployments, logs, env vars, domains.

Wolfram MCP Server

Official Wolfram MCP server for Wolfram Language evaluation and Wolfram|Alpha computational queries.

Zendesk MCP Server

Zendesk MCP — tickets, users, macros, knowledge base search for support agents.

Voice Infrastructure 10

STT, TTS, telephony for voice agents.

AssemblyAI

Speech-to-text and speech understanding API for developers building voice products.

Cartesia

Low-latency voice AI platform built around the Sonic text-to-speech model.

Deepgram

Voice AI platform providing speech-to-text, text-to-speech, and a unified voice agent API.

ElevenLabs

Voice AI platform offering text-to-speech, speech-to-text, and a conversational agents platform.

LiveKit Agents

Open-source realtime framework from LiveKit for building voice, video, and multimodal AI agents.

OpenAI Realtime API

Speech-to-speech realtime model API for building low-latency voice agents.

Pipecat

Open-source Python framework from Daily for building realtime voice and multimodal conversational agents.

Retell AI

Voice agent platform focused on low-latency conversation.

Vapi

Voice AI platform: STT, TTS, telephony, turn-taking for voice agents.

Whisper large-v3

Open-weights automatic speech recognition (STT) model.

Browser Automation 4

Headless browser, scraping, web actions.

Browser Use

Open-source framework that lets LLM agents control a real browser to complete web tasks.

Browserbase

Managed cloud browsers for agentic web automation.

Playwright

Microsoft's open-source browser automation framework, increasingly used as the execution layer for AI agents.

Stagehand

AI-native browser automation library; act/extract/observe primitives.

Payments / x402 4

Programmatic micropayments for agent-to-agent commerce.

Coinbase x402

HTTP 402 payment protocol for agent-to-agent and agent-to-API micropayments.

PayPal Agent Toolkit

PayPal's open-source toolkit and MCP server for integrating PayPal commerce APIs into AI agent workflows.

Skyfire

Identity and payments network for AI agents built on KYA (Know Your Agent) verification and token-based transactions.

Stripe Agent Toolkit

Stripe SDKs and MCP server that let AI agents create and manage Stripe payment objects via function calling.

Guardrails / Safety 5

Input/output filtering, jailbreak detection, PII redaction.

Guardrails AI

Open-source Python framework for validating and correcting LLM inputs and outputs using composable validators.

Invariant Labs

Guardrails and security analysis for agent tool use.

Lakera Guard

Commercial low-latency API for detecting prompt injection, data leakage, and unsafe content in LLM applications.

Meta Llama Guard

Family of open-weights LLM-based safety classifiers from Meta for moderating prompts and responses.

NVIDIA NeMo Guardrails

Open-source toolkit from NVIDIA for adding programmable guardrails to LLM-based conversational applications.

Model Routing / Gateway 4

Cost/latency-aware routing between models or providers.

LiteLLM

Open-source LLM gateway exposing 100+ providers through a single OpenAI-compatible API, as a Python SDK or proxy server.

OpenRouter

Unified API across 300+ LLMs from many providers; one API key.

Portkey

AI gateway: routing, caching, observability, guardrails across providers.

Requesty