Browse 57,964 MCP Servers

By mcp-dir

# Prefeitura MA São Luís: Certidão Negativa de Débitos Prefeitura MA São Luís: Certidão Negativa de Débitos, consulta em fonte oficial. Hospedado pela plataforma, sem credenciais da plataforma, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_pref_ma_sao_luis_cnd`

By accounts-ynho

TrueProxies MCP for residential and datacenter proxy account, traffic, and endpoints. Residential Unlimited is 15M+ IPs. No free trial. No Softylus. No UK.

By drbaher

Agent-first e-signature MCP server with **offline PAdES signing** and hash-chained audit. The same 19 tools / 4 prompts / 12 resources are reachable via stdio (`npx @drbaher/sign-cli mcp serve`) for local agents and over HTTP for remote ones. **Architectural claim:** an LLM agent can drive every step of a contract workflow except the actual signing gesture, which stays gated behind a human via per-signer approval tokens (TTL-bounded, scoped to one email, single-use). Pre-sign safety checks (`--require-hash`, `--require-title`, `--require-signer-email`) throw structured errors *before* any state mutation — so an agent that computed a hash earlier can refuse to sign if the document was swapped mid-flight. **Highlights:** - Fully-offline PAdES signer (real PKCS#7, self-issued X.509) - Hash-chained audit log with append-only DB triggers + RFC 3161 timestamping - Read-only mode (`--read-only true`) returns `FORBIDDEN_READ_ONLY` for mutating tools - Multi-provider routing (local / Dropbox Sign / DocuSign / SignWell) - Pre-sign signature visibility (`existingSignatures` on every fetch) - Tool allow-list + capability scoping for sandboxed agents This Smithery deployment runs in **read-only mode** against an ephemeral SQLite that wipes every 4 hours — safe to explore, not for production signing. For local agent use, install via `npm i -g @drbaher/sign-cli`. MIT-licensed. No telemetry. Source: github.com/DrBaher/sign-cli. Showcase: cli.drbaher.com.

By vittoria-builds

Market events for AI agents: 8-K filings, volume spikes, big day moves, opening gaps and upcoming earnings, over MCP. │

By hi-lmlo

CalmSEO is a hosted MCP server for SEO work. Connect over OAuth and query your Google Search Console properties for clicks, impressions, CTR, position, top queries, top pages, and sitemaps at no cost. Credit-based tools add live Google and Bing SERPs, keyword volume and CPC, keyword ideas, ranked keywords, and page audits.

By mcp-dir

# ECRVSP Restrições: Bloqueios Ativos DETRAN ECRVSP Restrições: Bloqueios Ativos DETRAN, consulta em fonte oficial. Hospedado pela plataforma, sem credenciais da plataforma, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_ecrvsp_restricoes_detran`

By mcp-dir

# SEFAZ CE: NFC-e SEFAZ CE: NFC-e, consulta em fonte oficial. Hospedado pela plataforma, sem credenciais da plataforma, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_sefaz_ce_nfce`

By formstep

Collect verified customer information with one request. Your agent sends a request to one customer, the recipient: a form in your brand and their language, prefilled with what you know, so they only confirm or correct it. They upload files, sign, book or pay on the same page, without an account. The answers come back under your field keys, with your external ID and, when the form has a decision question, an approve, decline or changes outcome. Your agent reads them back, or Formstep calls your callback URL the moment the recipient submits.

By buckhead-energy

Read Buckhead company information; prepare and send a mineral rights inquiry.

By a1-x-tech

MCP-сервер для Yandex Direct — управляйте контекстной рекламой (кампании, ставки, объявления, статистика) из Claude, Cursor, Codex и других AI-клиентов.

By hello-h7cc

Quantral's MCP server connects an AI assistant to Quantral's stock sentiment data: ask what retail investors and market commentators are saying about a stock, and get scores and monthly recaps back as interactive cards in the chat, with the mentions behind each score one question away. Five read-only tools: `get_top_signals` (strongest chatter right now, 24h or 7d window, ranked 0 to 100), `search_companies`, `get_company_score` (both windows with a tier word: Exceptional, Strong, Moderate, Weak, Negative), `get_company_recaps` (month-by-month summaries) and `get_company_signals` (the individual posts behind a score, up to 50 per request). The score is aggregated social sentiment from the sources Quantral tracks, weighted by source and author credibility. It is not a price prediction or a recommendation to buy or sell. Hosted, OAuth sign-in with a Quantral account (active subscription required), 60 requests a minute. Setup guide: https://quantral.com/mcp

By akash-agl92

MCP server for personal/mundane astrology grounded in statistical research. Discover evidence-based birth charts and daily transits that combine Western depth with time-tested Vedic life cycles, the Erux's unique Hybrid chart system with higher predictive accuracy. Receive clear, grounded guidance for your career, relationships, and personal growth — along with the exact best windows of the day to make important decisions, connect, or recharge.

By sadri

If-None-Match star-or-etag, value discarded

By mcp-dir

# Registro Civil: Validar Certidão Eletrônica Registro Civil: Validar Certidão Eletrônica, consulta em fonte oficial. Hospedado pela plataforma, sem credenciais da plataforma, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_registro_civil_val_cert_eletr`

By samuel-fraga

Overview Nevent AI is the official MCP server for Nevent, the marketing platform for live-event promoters — festivals, concert venues, clubs and sports organizations. Connect Claude, ChatGPT, Cursor or any MCP client to your event CRM and work with your real marketing data in natural language: query campaign performance, build audience segments, check deliverability, analyze cross-channel attribution — or go further and create segments, draft templates and schedule campaigns, with your approval before anything runs in production. What you can do Campaigns: query and manage email, SMS, WhatsApp and push campaigns — performance, funnels, scheduling Audience segmentation: explore attendee data and build segments by purchase behavior, event attendance or engagement Analytics & attribution: revenue per campaign, cross-channel attribution, event insights Deliverability: bounce, complaint and domain-reputation monitoring Templates & media: draft email templates and manage creative assets Paid media & short URLs: audience sync and tracked links Multi-tenant: manage several event brands from one connection

By agung212gumelar

Simceptor — AI-Native Mock API Server for Modern Development Teams Simceptor is a cloud-based mock API platform built from the ground up for AI-assisted development workflows. With 26 MCP tools, your AI coding assistant (Claude, Cursor, Windsurf, etc.) can create, configure, and manage your entire mock API layer — no dashboard clicking required. 🤖 Set Up Mock APIs Directly via MCP Simply tell your AI agent: "Create a POST /payments endpoint that returns a 200 success response." Simceptor's MCP integration handles the rest — from creating the endpoint to configuring response bodies, headers, and status codes, all in seconds through natural language. ⚡ Powerful Rule Engine Go beyond static mocks. Define conditional response rules based on request headers, body fields, query params, or path variables. Return different responses depending on the input — simulate edge cases, error states, and complex branching logic without writing a single line of server code. 🔄 Stateful API Simulation Simulate real-world stateful flows such as order lifecycles, authentication sequences, or multi-step checkout processes. Simceptor maintains session state across requests, letting you mock workflows that evolve over time — critical for end-to-end testing and demos. 📡 Webhook Support Configure your mock endpoints to fire outbound webhook calls upon being triggered. Test how your services react to external events without needing a real third-party system in the loop. 🧠 AI-Powered Issue Tracker When your mock doesn't behave as expected, raise an issue directly from the dashboard. Your AI agent can then read the issue, analyze the current mock configuration, apply a fix, and automatically mark it resolved — closing the feedback loop between humans and AI in your development cycle. Perfect for backend developers, QA engineers, and platform teams who want to move fast without waiting for real APIs to be ready. 🔗 simceptor.com

By wondercal

# WonderCal Schedule meetings that work for everyone — right from your AI assistant. WonderCal finds times that overlap across every participant's real calendar availability, so you skip the back-and-forth of "does Tuesday work?" ## What it does Tell your assistant who you're meeting with and it handles the whole loop: create the meeting, invite people to share availability, find the mutual overlap, and lock in a time — then drop it on everyone's calendar. ## Capabilities - **📅 Smart scheduling** — Create meetings and instantly surface times that work for all participants (`schedule-meeting`, `find-available-times`, `confirm-meeting`) - **🔗 Calendar sync** — Connect Google and Microsoft Outlook calendars for real-time free/busy (`connect-calendar`, `connect-microsoft-calendar`, `get-free-busy`) - **👥 Participants & invites** — Add people, send and track invitations, generate shareable invite links (`add-participant`, `list-invites`, `get-invite-link`) - **✏️ Full meeting control** — Update duration, title, or location; reschedule; set response deadlines; cancel — all while a meeting is unconfirmed (`update-meeting`, `reschedule-meeting`, `set-meeting-deadline`) - **📇 Contacts** — Look up and invite people you've met with before (`search-contacts`, `list-contacts`) - **🔗 Personal booking link** — Share a scheduling link so others can grab time with you (`get-scheduling-link`) ## Authentication Connect once with **Google sign-in** — no API keys to copy or paste. WonderCal handles OAuth securely and only ever reads your free/busy times. Learn more at [wondercal.com](https://wondercal.com)

By hasdata

Fetch any public web page through managed proxies, with optional JavaScript rendering, wait conditions and CSS or AI extraction rules, as markdown, text, HTML or JSON. Runs on HasData's hosted API.

By mcp-dir

# ECRVSP Processos: Andamento ECRVSP Processos: Andamento, consulta em fonte oficial. Hospedado pela plataforma, sem credenciais da plataforma, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_ecrvsp_processos_andamento`

By debtstack

Corporate credit data API for AI agents. Search bonds, leverage ratios, guarantors, corporate structure, and SEC filings across hundreds of companies. Screen high-yield bonds by YTM and seniority, resolve CUSIPs from free text, traverse guarantor hierarchies, and search full-text SEC documents.

By glenn-quinting

Price any x402 (HTTP 402 Payment Required) API across payment rails and see what the same call costs on each — live, from the server's own quote. Gas rails floor micro-prices at ~$0.001; the feeless Nano (XNO) rail quotes the true metered amount, so sub-millicent calls cost sub-millicents. **Tools:** `x402_quote` · `x402_compare` · `faucet_info` · `get_started` This hosted endpoint is read-only and holds no wallet — nothing to steal, nothing to spend. For the full self-custodied experience (pay per call, earn as a merchant, claim free starter XNO from the faucet): pip install feeless402 claude mcp add feeless402 -- uvx feeless402 mcp Docs: [feeless402.com](https://feeless402.com) · Spec: [railhint.com](https://railhint.com) · Official MCP registry: `com.feeless402/nano-pay`

By surfether

Four x402-paid domain verification tools for AI agents — trust scoring utilizing WHOIS, SSL certificate intelligence, HTTP security header audit, and robots.txt + AI bot policy detection. Pays per call in USDC on Base Mainnet. Each tool settles its own micropayment ($0.002–$0.003 USDC) atomically per call. No API keys, no signups — just a funded Base wallet. Suggested starter balance: $1 USDC + 0.0005 ETH covers ~300 calls. Tools are also discoverable to autonomous agents via Coinbase's Bazaar marketplace at agentic.market.

By billingengine

Create customers and invoices and record payments in <a href="https://www.billingengine.com">BillingEngine</a> from any MCP client. Invoices are created as drafts, never sent, and nothing can be deleted. BillingEngine is invoicing software for freelancers and small businesses with EU e-invoicing (EN 16931, XRechnung, ZUGFeRD / Factur-X). Requires a paid plan and the API key from Settings > API. Your key stays on your computer.

By manganate06

**[OpenLMNP](https://openlmnp.fr)** — free, open-source accounting for French furnished-rental hosts (Airbnb, Booking…), managed from your AI assistant, in plain language. Log income & expenses, track depreciation, compare tax options, and prepare your yearly tax return. 44 tools · works with Claude, ChatGPT, Gemini, Copilot & any MCP client. Try the free live demo of the **[MCP server for LMNP accounting →](https://openlmnp.fr/mcp)**. Self-hostable · AGPLv3.

By info-3fh4

OAuth MCP for LeanTrading Pro, Elite, Blueprint, Admin, and Product Tester (first 14 days of the tester grant). After that window, Pro is required. . Access strategy context, intelligence, risk preview, and TradingView Charting Agent tools from Cursor or Claude. Requires eligible paid account; TV chart actions need a paired ChartingBridge.

By hasdata

Search Airbnb stays by location and dates, and read a single listing in full, as structured JSON. Runs on HasData's hosted API.

By mcp-dir

# Prefeitura PE Recife: CPOM Prefeitura PE Recife: CPOM, consulta em fonte oficial. Hospedado pela plataforma, sem credenciais da plataforma, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_pref_pe_recife_cpom`

By goodturnstudio

Gurbani a few lines a day in Gurmukhi with transliteration and English, any Ang or shabad (English from BaniDB, credited), and whole banis such as Japji Sahib and Sukhmani Sahib read a passage at a time. Free and read-only: no sign-in, no key, and nothing about you is stored. Built by Good Turn Studio.

By nlqdb

Analytical memory for AI agents: a real Postgres your agent connects to over MCP and queries in plain English — GROUP BY, JOIN, aggregate over what it remembered, not just the top-k a vector store recalls. One command to connect.

By sadri

Current local time in Australia/Perth.

By sadri

CUID prefix-c shape, value discarded

By gabriel-naulleau

Western, Vedic, and Chinese astrology calculations, charts, forecasts, compatibility, BaZi, Panchang, and geocoding over hosted MCP.

By mailcheer

Send transactional email and newsletter campaigns from an agent: subscribers, tags, segments, suppression list, sending and stats. Connect with OAuth 2.1 or an API key. Runs on Amazon SES, hosted in Europe.

By mcp-dir

# Investigação Cadastral Investigação cadastral de pessoas e empresas: localização (endereço/telefone/email), vínculos e relacionamentos societários, participações, beneficiário final, renda estimada, óbito e propriedade veicular. Hospedado pela plataforma, sem credenciais, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_investigacao`

By mcp-dir

# SEFAZ TO: NFC-e SEFAZ TO: NFC-e, consulta em fonte oficial. Hospedado pela plataforma, sem credenciais da plataforma, pague por consulta com crédito pré-pago. Funciona em qualquer cliente MCP. Servidor remoto: `https://api.mcp.ai/p_sefaz_to_nfce`

By sadri

RFC 3339 UTC stamp

By coinfuty

Free, no-key MCP server for live crypto futures market data from [Coinfuty](https://www.coinfuty.com) — open interest, funding rates, long/short ratios, liquidations, volume and price across 25+ derivatives exchanges. **7 read-only tools:** market overview (total OI, 24h liquidations, Fear & Greed, average RSI) · coin screener · per-coin derivatives summary · historical series (price / OI / funding / long-short / liquidation / volume) · funding rates across exchanges · per-exchange breakdown · exchange list. - Endpoint: `https://mcp.coinfuty.com/api/mcp` (Streamable HTTP, no auth required) - Rate limit: 60 requests/min per IP - Docs & example prompts: [coinfuty.com/mcp](https://www.coinfuty.com/mcp)

By hello-la0s

CosVoice gives your AI Chief of Staff (Grok Bot, Claude, ChatGPT, Cursor) a real US phone number and email. It calls, answers, and sends written summaries. Connect via https://cosvoice.com/mcp — from $29/mo.

By chris-2dib

Built for newsrooms and AI agents that need to know not just when an election is but when its result will actually exist. Every stage window carries a confidence level and a citation to the statute or authority page it rests on. A result is only final after its final-result window — /v1/check and the MCP tool check_question answer whether a question is answerable yet, which is the part that stops an agent asserting a result that does not exist. Keyless API, 60 requests/minute. Corrections are public and take an evidence URL.

By nicolas-coutureau

FFmpeg as a service for AI agents. KinoPipe gives agents hosted video editing over remote MCP: 39 typed tools (trim, resize, crop, compress, convert, subtitles, watermark, GIF, scene and silence detection, local-file uploads) plus composable single-pass recipes with `create_video_job`. OAuth 2.1 sign-in, async jobs with `get_job`, signed download URLs. No shell commands: every request is validated before anything renders. Free plan: 100 credits, no credit card. Docs: https://kinopipe.com/docs. Connect guide: https://kinopipe.com/connect

By goodturnstudio

Talmud a few lines a day, explained line by line, with today's Daf Yomi, search, the original text of any page from Sefaria, and whole tractates in English read a passage at a time. Free and read-only: no sign-in, no key, and nothing about you is stored. Built by Good Turn Studio.

By gusmaogabriels

Read-only access to Gabriel S. Gusmao's research profile, publications, scientific machine learning playgrounds, citations, and local CLI discovery for KINN (PyPI: kinnlib), Optinpy, mkin4py, and xl2py. Eight tools provide verified links, installation instructions, examples, and software and methods-paper citations. Computations run locally on the user's computer; the remote MCP returns metadata only. Public Streamable HTTP endpoint; no API key required.

By cybergarden

Odoo ERP for AI agents: hosted OAuth endpoint, gated writes, one endpoint for every instance.

By evozim-hv

Rule-Porter is a premium micro-service from the M2MCent factory, configured as a Model Context Protocol (MCP) server. Functionality: Multi-IDE settings translator mapping shortcuts between apps. Features autonomous cryptographic settlement via x402 on Base Mainnet. Fee: $0.06 USDC per execution.

By bitfreak

# CAIT - Core AI Toolkit A modular [Model Context Protocol (MCP)](https://modelcontextprotocol.io) server that extends AI assistants with practical capabilities: file I/O, a persistent Python REPL, AST-aware code analysis, semantic text search, document conversion, Wikipedia & arXiv tools, a persistent vector memory database, and other general utilities. A total of **37 tools across 9 modules**. Each module can be disabled independently via the `CAIT_DISABLE` environment variable. Made by AI for AI. ## Requirements - Python 3.11+ - Core: `fastmcp`, `chromadb` - Online research: `wikipedia-api`, `arxiv` - Document conversion: `docling` or `markitdown` - Scientific computing (optional, for REPL use): `sympy`, `scipy`, `matplotlib`, `plotly`, `vispy` ## Installation Install the requirements into your Python environment then copy [copilot-instructions.md](https://github.com/JacobBruce/CAIT/instructions/copilot-instructions.md) or [CLAUDE.md](https://github.com/JacobBruce/CAIT/instructions/CLAUDE.md) into the correct location (more details below). The instructions include general guidance for how to behave, how to use CAIT tools, and how to use the Firecrawl search tools. The instructions may need to be adapted to suite different setups. If you are working in a Python environment you may want to make use of this agent prompt: [python-coder.agent.md](https://github.com/JacobBruce/CAIT/agents/python-coder.agent.md). There is also [research-assistant.agent.md](https://github.com/JacobBruce/CAIT/agents/research-assistant.agent.md) for deep research. There is also a skill file called [project-planning.md](https://github.com/JacobBruce/CAIT/skills/project-planning.md) which is helpful for planning the implementation details of a project. The agent produces a PLAN.md file and TASKS.md file. ### Environment Variables | Variable | Default | Description | |----------|---------|-------------| | `CAIT_FILES_PATH` | `~/.cait/files/` | Directory for downloaded files and document cache | | `CAIT_DISABLE` | _(empty)_ | Comma-separated module names to exclude at startup (e.g. `wiki,arxiv`) | ## Client Configuration ### VS Code (GitHub Copilot) Add to your workspace `.vscode/mcp.json` or user `settings.json`: ```json { "servers": { "bitfreak/cait": { "type": "stdio", "command": "/absolute/path/to/.venv/bin/python", "args": ["-m", "cait.server"], "cwd": "/absolute/path/to/CAIT" } } } ``` > For user `settings.json`, nest the above under `"mcp": { ... }`. Copy [copilot-instructions.md](https://github.com/JacobBruce/CAIT/instructions/copilot-instructions.md) into your project's `.github/` folder to give Copilot guidance on using CAIT tools effectively. ### Claude Desktop Edit `claude_desktop_config.json`: - **macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json` - **Linux**: `~/.config/claude/claude_desktop_config.json` - **Windows**: `%APPDATA%\Claude\claude_desktop_config.json` ```json { "mcpServers": { "bitfreak/cait": { "command": "/absolute/path/to/.venv/bin/python", "args": ["-m", "cait.server"], "env": { "PYTHONPATH": "/absolute/path/to/CAIT" } } } } ``` ### Claude Code ```bash claude mcp add cait -e PYTHONPATH=/absolute/path/to/CAIT \ -- /absolute/path/to/.venv/bin/python -m cait.server ``` Copy [CLAUDE.md](https://github.com/JacobBruce/CAIT/instructions/CLAUDE.md) to your project root (or to `~/.claude/CLAUDE.md` for global use) to give Claude Code guidance on using CAIT tools effectively. ## Recommended MCP Servers ### Firecrawl [Firecrawl](https://firecrawl.dev) is a web scraping and search API that pairs naturally with CAIT, adding powerful web search, full-page scraping, and site crawling. A free API key is available at [firecrawl.dev](https://firecrawl.dev). ### Serena [Serena](https://github.com/oraios/serena) provides many tools for semantic code retrieval and editing. Both CAIT and Serena include a similar memory system so it is recommended to disable one of them. ## Tool Reference ### File System — `fs` | Tool | Description | |------|-------------| | `get_file_info` | Metadata for a single file: size, line count, permissions, timestamps. Does not read content. | | `get_dir_info` | Directory listing with per-entry metadata. Supports glob patterns and recursion. | | `append_file` | Append text to a file. Useful for NOTES.md, TASKS.md, log files. | | `download_file` | Download a URL to `~/.cait/files/` (or `CAIT_FILES_PATH`). Returns the local path. | | `fetch_url` | HTTP GET/POST with custom headers and body. Use `save_to` to avoid large responses in context. `convert=True` returns clean markdown via Docling or MarkItDown. | ### Persistent Python REPL — `repl` | Tool | Description | |------|-------------| | `repl_exec` | Execute Python code in a persistent session. Variables, imports, and function definitions survive between calls. Returns stdout, stderr, and exception info. | | `repl_read` | Inspect a named variable from the REPL session without executing code. Returns repr, type, and JSON value. | | `repl_vars` | List all user-defined variables in the current REPL session. Returns name, type, repr, and JSON value for each. Useful for reviewing session state without running code. | | `repl_reset` | Clear all variables and imports from the REPL session. | ### Code Analysis — `code` All code tools perform **AST-aware** search — they skip occurrences in comments and strings, unlike text grep. | Tool | Description | |------|-------------| | `find_definitions` | Find all definitions of a function, class, or variable. Returns file, line, docstring, and kind. | | `find_calls` | Find all call sites of a function. Matches bare calls, method calls, and chained calls. | | `find_imports` | Find all files that import a given module or name. | | `find_references` | Find all uses of an identifier (loads, stores, deletes, attribute accesses). | ### Text Search & Embeddings — `text` Uses `all-MiniLM-L6-v2` (bundled with ChromaDB — no separate download). Chunk embeddings are cached in memory so repeated queries on the same document skip re-embedding. | Tool | Description | |------|-------------| | `search_text` | Semantically search or summarize a text string or plain text file (`.txt`, `.md`, `.rst`). Query given → extract mode (most relevant chunks). Query empty → summarize mode (most representative chunks). | | `encode_text` | Return raw 384-dimensional float embeddings for one or more strings or files. | | `text_similarity` | Cosine similarity between two texts (0–1). | ### Document Tools — `document` | Tool | Description | |------|-------------| | `convert_doc` | Convert PDF, DOCX, PPTX, XLSX, HTML, LaTeX, images, audio, and more to markdown or plain text. Backends: `docling` (higher quality, layout-aware), `markitdown` (lighter, better for Office files), `auto` (tries docling, falls back to markitdown). Use `save_to` to write large outputs to a file. `strip_tables=True` removes noisy pipe-table syntax. `rich_pdf=True` enables Docling's code detection and formula extraction (slower). | | `search_doc` | Same as `search_text` but handles many document formats (PDF, DOCX, HTML, URLs). Converts via `convert_doc` on first call and caches the result — repeat calls are instant. | ### Wikipedia — `wiki` | Tool | Description | |------|-------------| | `wiki_search` | Search Wikipedia. Returns titles, snippets, word counts, and URLs. | | `wiki_sections` | List all sections of a page as a table of contents (no text). | | `wiki_section` | Get the text of a specific section. Use `wiki_sections` first to find section titles. | | `wiki_page` | Get full page text or just the summary (`summary_only=True`). Supports non-English via `language` parameter. | ### arXiv — `arxiv` | Tool | Description | |------|-------------| | `arxiv_search` | Search arXiv. Supports field prefixes (`ti:`, `au:`, `abs:`, `cat:`) and boolean operators. Returns metadata for up to 100 papers. | | `arxiv_paper` | Fetch a paper by ID. `full_text=False` (default) returns abstract + metadata. `full_text=True` downloads and converts the full PDF. Use `save_to` for large outputs. | ### Datetime & Utilities — `utils` | Tool | Description | |------|-------------| | `get_datetime` | Current date, time, timezone, UTC offset, weekday, and Unix timestamp. Accepts any IANA timezone name. | | `timer_start` | Start a named wall-clock timer. | | `timer_stop` | Stop a timer and return elapsed seconds. | | `timer_list` | List all running timers and their current elapsed time. | | `diff_text` | Unified diff between two strings or files. Returns diff text plus added/removed line counts. | ### Memory Database — `memory` Persistent ChromaDB vector store at `~/.cait/files/` (shared across projects). Content is embedded with `all-MiniLM-L6-v2` for semantic retrieval. | Tool | Description | |------|-------------| | `mem_add` | Add a new entry. Fields: `title`, `content` (embedded), `tags`, `description`, `source`, `entry_id`. | | `mem_search` | Find entries by semantic similarity to a query. Optionally filter by tags. | | `mem_get` | Retrieve a full entry by ID. | | `mem_list` | List entries sorted by date (newest first). Content omitted for brevity. | | `mem_set` | Update fields of an existing entry. Only non-empty values are applied. | | `mem_edit` | Edits content in-place — regex replace when pattern is given, or append when not. | | `mem_delete` | Permanently delete an entry by ID. | | `mem_find` | Fast metadata scan — no embedding. Match by title substring, exact source URL, or tags. Use this for deduplication checks before `mem_add`. | ## Disabling Modules Set `CAIT_DISABLE` to a comma-separated list of module names to exclude their tools at startup: ```bash CAIT_DISABLE=wiki,arxiv python -m cait.server ``` Available module names: `fs`, `text`, `code`, `repl`, `wiki`, `arxiv`, `utils`, `memory`, `document`

By calvinling2021

**The GEO check every vibe coder runs before they launch.** Synapse is the GEO (Generative Engine Optimization) growth layer for new products. This MCP server gives any coding agent — Cursor, Claude Code, Windsurf, Cline, Continue, Zed — the ability to audit, fix, and track how AI agents discover and recommend a project. ### Tools - **`geo_check`** — Run the 24-rule GEO linter on a URL or local path. Returns a 0–100 Growth Score and a list of failing rules with evidence. - **`geo_fix`** — Apply auto-fixers: writes `llms.txt`, `robots.txt` (with the AI-bot allowlist), and `/.well-known/agent-answer.json` to the project. Never touches your content. - **`geo_track_init`** — Register the site with the Synapse dashboard to start tracking AI-agent referrals. - **`geo_prompts`** — Generate a visibility-test prompt pack for ChatGPT / Claude / Perplexity. - **`geo_status`** — Live Growth Score, agent mentions, and visits for an installed site. - **`geo_corpus_query`** — Query the public Synapse corpus to find products matching a user intent. ### Resources - `synapse://docs/llms-txt-spec` — the canonical llms.txt reference - `synapse://docs/geo-checklist` — the full 24-rule rubric - `synapse://corpus/latest` — recent agent-recommendation tuples ### Install — Cursor / Claude Code / Windsurf / Cline / Continue / Zed ```json { "mcpServers": { "synapse-geo": { "command": "npx", "args": ["-y", "@calvin8miles/mcp-server"] } } }

By ply-ui-ng

List, search, and inspect Ply Angular + Tailwind copy-in components. Hosted catalog at mcp.ply-ui.com. Local stdio (npx -y ply-ui-mcp) can also add, init, doctor, diff, and update.

By sadri

Current local time in America/Phoenix.