Best AI & LLM Tools MCP servers for Codex

Lets an AI client run browser tasks in a real Chrome browser, including fills forms, clicks, extracts data, and downloads files.

[mcp_servers.browser-use]
command = "uvx"
args = ["browser-use"]

An open-source AI agent framework that enables running multi-agent teams and workflows locally or via cloud providers.

[mcp_servers.praisonai]
command = "uvx"
args = ["praisonai"]

Fetches up-to-date, version-specific library documentation and code examples into an AI coding agent's context.

[mcp_servers.context7]
command = "npx"
args = ["-y", "@upstash/context7-mcp"]

Lets AI clients search n8n nodes, templates and docs, validate configurations, and build or deploy n8n workflows.

[mcp_servers.n8n-mcp]
command = "npx"
args = ["-y", "n8n-mcp"]

Mocks AI infrastructure including LLM APIs, MCP tools, vector databases, and agent protocols for deterministic application testing.

[mcp_servers.aimock]
command = "npx"
args = ["-y", "@copilotkit/aimock"]

Provides AI assistants with real-time web search, reasoning, and deep research capabilities using the Perplexity API.

[mcp_servers.modelcontextprotocol]
command = "npx"
args = ["-y", "@cormac315/perplexity-mcp"]

Turns web pages into clean markdown, JSON, or LLM-ready text and lets AI clients scrape, crawl, search, and extract structured data from websites.

[mcp_servers.webclaw]
command = "npx"
args = ["-y", "@webclaw/mcp"]

Keeps bulky tool output out of a coding agent's context window by running work in a local sandbox and indexing results in SQLite for on-demand search.

[mcp_servers.context-mode]
command = "npx"
args = ["-y", "context-mode"]

Lets AI clients search the web, fetch full page content as markdown, and run multi-step research agents.

[mcp_servers.exa-mcp-server]
url = "https://mcp.exa.ai/mcp"

Lets an AI client orchestrate swarms of specialized agents, coordinate workflows, and retain self-learning memory across sessions.

[mcp_servers.ruflo]
command = "npx"
args = ["-y", "@claude-flow/cli-core"]

Enables AI agents to discover, recommend, and select optimal orchestration frameworks and agentic harnesses based on specific models and tasks.

[mcp_servers.best-of-agent-harnesses]
command = "uvx"
args = ["agent-harnesses-mcp"]

Lets an AI client search, inspect, lint, and install Agent Skills (SKILL.md) from the SkillMD registry.

[mcp_servers.skillmd]
command = "npx"
args = ["-y", "get-skill"]

Lets an AI client search and discover Agent Skills from the skills.sh registry to extend agent capabilities.

[mcp_servers.agentskills]
url = "https://skills-sh.run.mcp.com.ai/mcp"

Lets an AI client mock, record, replay, verify and chaos-test HTTP, gRPC and LLM API dependencies.

[mcp_servers.mockserver-monorepo]
command = "docker"
args = ["run", "-i", "--rm", "docker.io/mockserver/mockserver:9.2.0"]

Lets an AI client run Google searches and return structured JSON with organic results, AI Overview, People Also Ask, news, and shopping data.

[mcp_servers.google-search-mcp]
command = "npx"
args = ["-y", "@hasdata/google-search-mcp"]

Provides a two-layer memory system for AI agents with an immune system to validate and consolidate episodic experiences into long-term identity.

[mcp_servers.anneal-memory]
command = "uvx"
args = ["anneal-memory"]

Generates architecture diagrams, explanations, Mermaid source, and narrated videos for GitHub repositories.

[mcp_servers.gitdiagram]
url = "https://gitdiagram.com/mcp"

Provides memory infrastructure for stateful agents to store, reason over, and query conversations, events, and peer representations.

[mcp_servers.honcho]
url = "https://mcp.honcho.dev/"

Provides an ecosystem of over 2,700 scientific tools, datasets, and machine learning models for AI agents to perform research and data analysis.

[mcp_servers.tooluniverse]
command = "uvx"
args = ["tooluniverse"]