Agent tools

Web data API designed for AI agents and developers to search, scrape, crawl, and extract structured data from websites at scale, returning clean Markdown and JSON ready for retrieval and downstream processing.

Open-source library enabling AI agents to operate web browsers like humans, providing automated browsing and task execution for web automation workflows.

Persistent memory system for coding agents that captures session activity, compresses it with AI, and injects relevant context into future sessions across Claude Code, Codex, Gemini, and others.

Open-source Retrieval-Augmented Generation engine combining document parsing, retrieval, and agent capabilities to provide a context layer for enterprise LLM applications.

Lets AI agents read and search content across Twitter, Reddit, YouTube, GitHub, Bilibili, and Xiaohongshu through one CLI, providing developers and researchers broad web access without paying platform API fees.

Offers an open-source, large-language-model-friendly web crawler and scraper that extracts clean, structured data for retrieval-augmented generation and autonomous agent pipelines.

Context compression layer for AI agents that reduces tool outputs, logs, files, and RAG chunks before inference, cutting token usage while preserving answers via library, proxy, and MCP.

Local-first developer tool that pre-indexes codebases into a knowledge graph with semantic search, helping AI coding agents use fewer tokens and tool calls across multiple IDEs.

Provides drop-in memory infrastructure for AI agents and applications, maintaining persistent cross-session context and personalized recall suitable for production deployments.

Local-first open-source memory system for AI applications, organizing stored conversation content around people and projects with semantic search and pluggable vector backends.

Collects real-time LLM data-pipeline templates built on Pathway, demonstrating high-accuracy retrieval-augmented generation and enterprise search-style question answering over live data sources.

Lets coding agents control and inspect real Chrome browsers through an official MCP server, enabling automated debugging, testing, and performance analysis for web developers.

Framework making arbitrary professional software agent-native by generating CLIs, providing a registry hub, and enabling trajectory replay for automation for AI developers.

Builds interactive code knowledge graphs entirely in the browser from Git repositories or ZIP uploads, giving developers Graph RAG question answering and MCP tools for exploring dependencies across large codebases.

A high-performance local code-intelligence service that indexes codebases into a persistent knowledge graph with structured retrieval, designed to reduce token usage for coding agents.

Gives AI agents a dedicated browser-automation CLI using accessibility-tree references to navigate pages, click elements, fill forms, capture screenshots, and perform repeatable web operations programmatically.

Builds a self-evolving context database for AI agents, unifying memory, RAG knowledge, and skills. Exposed as a virtual filesystem, it handles storage, retrieval, and accumulation of conversational memory.

MCP server exposing Playwright browser automation to large language models through structured accessibility snapshots. Enables AI agents to navigate, read, and operate web pages without vision-based input.

Modular retrieval-augmented generation system that converts unstructured documents into graph-based indexes for question answering. Helps developers and researchers query large text collections with structured, community-aware retrieval.

Indexes long documents into hierarchical page structures and uses LLM reasoning for vectorless retrieval, offering developers an alternative to embedding-based RAG.

Local-first code intelligence graph that builds persistent repository maps for MCP servers and CLIs, letting AI tools fetch relevant context on demand and reduce token consumption.

An AI-based Python scraping library that turns natural-language prompts into extraction pipelines using graph logic to produce structured data from webpages.

Enables AI agents to read, edit, and automate Word, Excel, and PowerPoint files through a free open-source single binary requiring no Office installation, with document rendering and preview support.

A framework for building real-time temporal knowledge graphs that give AI agents evolving contextual memory with hybrid retrieval and historical queries.

Open-source AI memory platform providing persistent cross-session long-term memory for agents through a self-hosted knowledge graph engine with multi-format ingestion and retrieval.

A tool integration layer for AI agents offering search across more than a thousand application toolkits with authentication, context management, and sandboxed execution.

Adds controllable shared memory for OpenClaw and Hermes agents by storing sourced facts in a knowledge graph, supporting corrections, decay, and reuse so multiple agents can retain consistent long-term context.

A local-first command-line search engine for Markdown docs, knowledge bases, and meeting notes, combining full-text search, vector retrieval, and LLM reranking with MCP access for agents.

Equips coding agents with semantic code retrieval, editing, refactoring, and debugging capabilities through MCP tools. Often described as an IDE for agents, it supports autonomous software engineering workflows.

A hub that turns websites and local tools into a unified command line, letting AI agents operate logged-in browsers to automate tasks for power users.

A community Model Context Protocol plugin connecting large language models to Blender, enabling creators to generate assets by creating objects, materials, and scenes or executing code.

Lightweight in-page GUI agent running entirely in JavaScript that lets users control web interfaces with natural language for tasks such as forms and navigation.

Persistent memory system for AI coding agents built on the iii engine, enabling cross-client storage, retrieval, and reuse of memories with real-world benchmark validation.

Lightweight headless browser engine written in Rust, compatible with CDP and Puppeteer/Playwright for webpage rendering and scraping by AI agents and crawlers.

Provides distributed graph-based task memory for coding agents, tracking dependencies, synchronizing state across workers, and persisting long-term recall across development sessions.

Offers a team-level memory hub for AI agents that converts conversations, documents, and code into governed, shareable assets such as chat memory, skills, wikis, and code graphs.

Open-source computer-use toolkit giving AI agents cross-OS desktop drivers, cloud-hosted isolated desktop clusters, local virtual machines, and evaluation benchmarks for automation research.

Long-term memory system for AI agents supporting retention, recall, reflection, and knowledge-base modeling, designed to sustain performance on extended tasks.

A Python library that lets analysts query SQL databases, CSV files, and other datasets in natural language via large language models, returning analytical results and visualizations.

Multimodal retrieval-augmented generation framework that parses text, images, tables, and formulas from documents, enabling developers to build enhanced retrieval and question-answering workflows over complex files.

An MCP server that exposes n8n node documentation and workflow-building capabilities to AI assistants, bridging the n8n automation platform with large language models.

Privacy-first tool stripping AI watermarks from user-owned content, removing invisible statistical signals and metadata provenance markers from text and multi-format files for agents and owners.

Local-first tool that continuously records screen and audio, applies OCR and transcription for semantic search, and supplies long-term visual context and memory for AI agents.

Unofficial NotebookLM programming interface and agent skill enabling AI agents to perform grounded research, content generation, and persistent memory management through outsourcing.

Open-source GenBI engine that provides AI agents with a governed semantic layer and context, helping data teams generate accurate SQL and deploy shareable dashboards.

Persistent memory infrastructure for AI agents that converts conversations and executions into structured state, helping enterprises integrate with existing data stacks across multiple deployment options.

Serverless single-file memory layer for AI agents that packages embeddings and retrieval indexes together, enabling developers to add instant recall and portable offline long-term memory.

Open-source MCP server that connects AI agents directly to enterprise databases, offering developers prebuilt generic tools plus a secure framework for defining custom database tools.

A browser built for humans and AI agents to share logged-in sessions side by side, enabling low-token web automation without setup, cost, or interrupting the user.

MCP server for Xiaohongshu that lets AI assistants search notes, retrieve content, and access engagement data, enabling programmatic access to the platform.

MCP server that supplies Figma design data to AI coding tools like Cursor, condensing layout and style information so models can generate accurate frontend code.

Lightweight memory system for coding agents that distills reusable skills from history and shares them across sessions with retrieval for reuse.

MCP service bridging AI assistants with the Unity editor to manage assets, manipulate scenes, edit scripts, and automate game development workflows.

Graph-native infrastructure for contextual AI systems, building situational knowledge graphs that support explainable causal reasoning and retrieval for AI engineers.

A portable, local-first memory layer for AI agents that stores user-owned knowledge in Markdown, synchronizes indexes across tools, and enables evolving reuse across workflows.

MCP plugin providing semantic code search for coding agents like Claude Code, indexing entire repositories into a vector store for on-demand retrieval as context.

MCP server built as a Chrome extension that exposes local browser capabilities to AI assistants, enabling automated control, navigation, and content analysis for agent developers.

MCP server bridging IDA Pro and large language models, letting AI coding assistants query disassembly, rename symbols, and add comments for reverse engineers.

MCP server connecting AI assistants like Claude to over 150 security tools, enabling automated penetration testing, vulnerability discovery, and bug bounty workflows for security professionals.

API service converting URLs to Markdown and providing web search for LLMs and agents, making fresh web knowledge easily accessible for RAG systems.

Python library that exposes existing FastAPI endpoints as authenticated MCP tools, allowing AI agents to discover and call current APIs directly for developers.

MCP server for coding CLIs like Claude Code enabling multi-model collaboration through cross-model consultation, consensus debate, and sub-agent orchestration for developers.

Long-term memory operating system for LLMs and agents that unifies storage, retrieval, and management of multimodal memories and knowledge bases, enabling hybrid search and cross-task skill reuse.

Anti-detect headless browser service for AI agents, built on a fingerprint-spoofed Firefox fork with Playwright compatibility, offering snapshots and automation interfaces.

High-speed file path and content search SDK with in-memory indexing, background watching, and MCP service for AI agents and editors.

Pixel-native visual retrieval engine for developers building LLM search, rendering web pages as screenshots, indexing appearance, and answering questions by visual content rather than parsed text.

Suite of AWS-specific MCP servers giving AI agents capabilities for documentation lookup, infrastructure deployment, data handling, and AI operations on AWS.

MCP server giving AI assistants such as Claude terminal control, file search, and diff-based editing, allowing agents to operate local files and execute commands for developer automation workflows.

TypeScript toolkit giving AI agents autonomous web browsing via Playwright, supporting clicks, input, navigation, and extraction with multi-model support and sandboxed execution.

Collects official Google Cloud Knowledge Catalog tools and samples demonstrating metadata management, context enrichment, and retrieval patterns for developers integrating enterprise knowledge and search workflows.

A high-accuracy RAG toolkit for scientific literature that parses and indexes PDF text, then uses LLM agents to answer questions with grounded, inline-cited responses for researchers.

A knowledge-augmented generation framework built on OpenSPG and large models, supporting domain knowledge-graph construction and logic-guided hybrid retrieval reasoning for specialized question answering.

Provides free remote MCP servers for any GitHub repository, allowing AI assistants to retrieve current documentation and source code directly to reduce hallucinations during coding assistance.

Provides long-term shared memory for AI coding assistants, enabling retrieval, persistence and handoff of project context across different agent vendors, tools and machines for development teams.

Provides a production-oriented agentic retrieval-augmented generation system centered on RESTful APIs, offering multimodal ingestion, hybrid search, knowledge graphs, document management, and deep research capabilities.

Open-source browser API and session infrastructure for AI agents, supporting web automation, content extraction, screenshots, PDF generation and CLI-based invocation.

Official MCP server exposing web search, scraping, interaction, and deep research to LLM clients, giving AI agents clean structured web context for retrieval-augmented workflows and autonomous browsing.

MCP server giving coding agents live browser observability by streaming console logs, network traffic, screenshots and Lighthouse audits from a real Chrome session for debugging.

MCP service and lightweight open-source toolset for Windows that lets AI agents operate files, applications, and UI elements, enabling computer automation for developers building desktop agents.

MCP service connecting AI agents to Figma, allowing agents to directly read design files and programmatically create or modify elements.

A codebase intelligence tool for developers and AI agents, providing health scores, automatic documentation, Git analytics, dead-code detection, and architectural insights through MCP access.

Enables agent-driven mobile automation and scraping across iOS, Android, emulators, and real devices through accessibility trees and coordinate-based actions over MCP.

Browser automation tool consisting of a CLI and extension that lets AI agents operate the user's real logged-in browser in isolated tasks without disrupting ongoing manual work.

Gives coding agents persistent, searchable local memory through an agent-agnostic Go binary backed by SQLite and FTS5, exposed over MCP, HTTP API, CLI, and TUI.

AI memory management system that emulates human brain encoding, consolidation, and pruning to provide dynamically evolving, human-like memory capabilities for agents.

Claude Skill that gives coding agents control of Chrome through persistent tabs, page snapshots, and result verification for browser automation during development and testing tasks.

MCP bridge connecting AI assistants with the local TradingView desktop application, enabling chart analysis, workflow automation, and strategy-related interactions for individual traders.

MCP service and CLI tooling for iOS and macOS development that enables AI coding agents to build, test, and manage Xcode projects, streamlining automated mobile development workflows.

Offers a unified real-time search skill for AI agents, supporting general web search, vertical domain search, parallel batch queries, and full-page content extraction for agent developers.

CLI plus MCP tool letting AI agents reuse the user's Chrome login state to control the browser, covering dozens of platforms and commands for search and web automation tasks.

A programmatic bridge for Google NotebookLM providing CLI, MCP server, and agent skill wrappers so AI agents can query and reuse notebook knowledge through commands.

Offers programmatic access to Gemini Notebook through a command-line interface, MCP server, and agent skills for retrieving and creating notebook content.

A fast local code search engine built for agents that indexes repositories and returns relevant snippets in milliseconds while using far fewer tokens than grep.

Browser automation CLI designed for AI agents, handling anti-bot bypass, human handoff, parallel multitasking, isolated multi-session browsing, and multi-account support across cloud or local execution.

An MCP server for Atlassian Confluence and Jira enabling AI assistants to search, create, and update tickets and documentation programmatically.

Brings AI-generated UI components and design inspiration directly into Cursor, Windsurf, and similar editors over MCP, helping frontend developers reuse code faster.

Extensible MCP integration platform for AI agents offering 100+ ready-made integrations, OAuth handling, sandboxed execution, and cloud plus SDK deployment for developers adding external tools.

Provides an MCP server for game developers and AI assistants working with Godot, enabling tools to launch the editor, run projects, and capture debug output automatically.

An MCP server built on Playwright that lets AI assistants operate real browsers to automate testing, web interaction, scraping, and API workflows in desktop IDEs.

Lets large language models create and edit flowcharts and architecture diagrams in Draw.io through tools, serving developers automating documentation and system design visuals.