MCP servers that ship as code — a repository or a package you install and run locally (usually over stdio). They have no hosted endpoint, so we cannot probe or grade them: instead each one shows its repository, install command, license, stars and when it was last updated.
Collected from GitHub repositories and npm, PyPI, Cargo, Go, Maven, NuGet and RubyGems packages tagged as MCP servers, the official MCP Registry, Docker's MCP catalog, ModelScope, Baidu MCP World, Cline and other public directories; one entry per repository. A server that also offers a hosted endpoint is listed under Remote instead. These entries are not counted in our reachability, verification or readiness statistics.
-
MCP server for time-travel kernel debugging: record a QEMU run, replay it, step backwards, and ask 'which instruction last wrote this address?' - with file:line.
MIT -
Python FastAPI recorder for MCP session replay, approval history, evidence chains, and operator-facing audit posture.
MIT -
@replay-build/mcp-serverupdated this year
MCP server for Replay.build — generate React UI from video recordings via AI agents
npm MIT -
@anyforgex/recordforgeactive
Record browser workflows over CDP and replay them deterministically. An MCP server + CLI that gives any agent-driven browser automation a capture-and-replay capability.
npm MIT -
A concurrent browser MCP server with session recording & replay - supports multiple parallel browser instances and captures all actions as JSON
npm Apache-2.0 -
agent-replay-mcpgithub.com/mdfifty50-boop/agent-replay-mcpupdated this year
MCP server for agent session recording and replay — debug non-deterministic agent behavior with session comparison and divergence detection
npm -
The fastest browser MCP server for AI agents: drive a real Chrome live, record a task once, replay it in milliseconds without an LLM, or replay the site's own API calls.
npm MIT ★ 1 -
Deterministic record/replay & time-travel debugging for AI agents. Local-first, MCP-native.
MIT ★ 1 -
Record browser workflows over CDP and replay them deterministically — an MCP server + CLI for agent-driven browser automation.
AGPL-3.0 ★ 1 -
🌐Web Agent Protocol (WAP) - Record and replay user interactions in the browser with MCP support
MIT ★ 508 -
MCP server for rr reverse debugging — enables AI assistants to record, replay, and inspect program executions via GDB/MI
pypi Apache-2.0 ★ 3 -
Record/replay HTTP mocking for fetch & axios — React Native/Expo-first, with scenarios, a dashboard, and an MCP server for AI assistants.
MIT ★ 2 -
AI-first mock server and HTTPS proxy: point a phone, simulator or any HTTP client at it to inspect, record, mock and replay API calls — driven by your coding agent over MCP or the CLI, or by code that embeds it.
npm MIT -
Let LLMs and AI agents "watch" video. Zakira.Replay turns any video source — YouTube URLs, conference recordings, course lectures, local files — into speaker-attributed transcripts (captions or local Whisper STT), frame-level OCR and structured vision (local or LLM-routed), slide-aware grouping, and timestamped JSON evidence that agents can reason over. Ships as both a CLI (`zakira-replay`) and an MCP server (`zakira-replay mcp serve`) in one .NET 10 binary.
nuget MIT -
Flight recorder for Cloudflare Workers: capture failed requests to .cfreplay.json and replay offline
npm MIT ★ 1 -
Replay the crash. Prove the fix. 81ns in-process flight recorder, 61µs deterministic replay, and 5.9ms variant fix verification for journal-driven services and AI coding agents.
go Apache-2.0 ★ 2 -
A local black-box recorder for AI agents: capture, replay, compare, and investigate failures.
NOASSERTION -
io.github.RudrenduPaul/agent-observabilitygithub.com/rudrendupaul/agent-observabilityno update data
AI agent observability with deterministic record/replay for debugging agent failures.
pypi -
Hash-chained audit ledger for AI agent work — record, verify and replay what the agent did.
npm -
AI-native MITM proxy — MCP server for traffic interception, recording & replay
npm Apache-2.0 ★ 19 -
Terminal automation for AI agents: PTY sessions, image-aware screenshots, visual regression, recording/replay, and MCP.
MIT ★ 1 -
witnessgithub.com/pmilet/witnessupdated this year
An MCP server that gives AI agents the power to record, replay, and compare HTTP API interactions — the REST equivalent of Playwright MCP.
Apache-2.0 -
lucarne-corpus-mcpgithub.com/volter-ai/lucarneupdated this year
Optional thin stdio MCP bin: read-only queries over a lucarne-records store. Never fetches, replays, or opens a browser — a miss returns a structured not_captured result telling the agent to browse to it in-session.
npm MIT -
把「重复点几十下」录一次,之后一键重放、定时跑,还能让 AI 通过 MCP 调用的浏览器操作录制器(Chrome MV3 扩展,Vue 3 + Vite)。 | Record repetitive browser actions once, then replay with one click, schedule them, or let AI drive them via MCP.
MIT -
tetris-mcpgithub.com/freezing/tetris-mcpupdated this year
A Tetris engine exposed as an MCP server — letting Claude play Tetris via structured tool calls. Built in Rust. Records full game history for replay and strategy analysis.
-
Flight recorder, MAST failure attribution & on-call ops loop for AI agents: replay, tamper-evident evidence, triage queue, letter grades, handoff briefs, fleet tail, price catalog, conversation recall. 53 MCP tools. Zero dependencies, Python 3.10+.
MIT ★ 1 -
Time-travel debugging for AI agents: record, fork and replay runs, trajectory evaluation, MCP server
-
obs-studio-mcpgithub.com/aaronckj/obs-studio-mcpupdated this year
MCP server for controlling OBS Studio over obs-websocket v5: scenes, sources, audio, streaming, recording, replay buffer, stats, and text overlays.
pypi MIT -
langbot-mcpactive
Langbot MCP server: every market on Kalshi and Polymarket, its history and replays against recorded order books, for Claude Code, Cursor, Claude Desktop and Codex.
pypi MIT -
vertex-mcp-geminigithub.com/michaelandrewgamble/vertex-mcp-geminiupdated this year
Minimal MCP gateway connecting Claude Desktop to enterprise AI providers using ADC authentication, with reproducible call recording and replay.
MIT -
AI-powered pentesting automation with MCP integration for Claude Code - parallel scanning, OWASP ASVS 5.0, workflow recording/replay
NOASSERTION ★ 1 -
Open-source E2E testing an AI agent records for you. Plain-language tests, recorded once on your real app (web, desktop, SAP, Citrix, AI agents), replayed on every commit with zero LLM calls.
npm Apache-2.0 ★ 6 -
Rewind CLI + stdio MCP server: whole-workspace checkpoint/rewind, a refuse-and-record effect barrier on a tamper-evident chain, and a token-saving record/replay proxy — for Claude Code, Cursor, and Codex CLI. Run with `npx @agent-rewind/mcp`.
npm FSL-1.1-ALv2 -
underpixelgithub.com/Leluth/underpixelupdated this year
Chrome extension + MCP server — record, replay, and understand what's behind the pixels. Timestamped visual-API correlation for Claude Code and any MCP client.
MIT -
Market-data engine for crypto and US equities: collect typed records, query a local Parquet lake, and replay observations through CLI, REST, or MCP.
Apache-2.0 ★ 4 -
Record once, replay with zero LLM tokens: a Chromium/CDP browser-control MCP server and CLI with deterministic skills, human login handoff, privacy de-identification and a hash-chained audit log.
Apache-2.0 -
A shared screen map and verified-workflow store for computer-use agents. Records observations on a normalized grid, connects screens with verified transitions, and replays a route with no model in the loop.
MIT -
Agent-friendly toolkit for debugging Sega Saturn recompilations: decode, render and diff video state, compare against reference emulators, and replay recorded play. CLI, MCP server and agent skills. No game data included.
MIT ★ 1 -
forgedgithub.com/sam-siavoshian/forgedupdated this year
Browser automation that learns. An MCP server that makes AI agents faster at repeated browser tasks by recording what works and replaying deterministic steps via Playwright.
MIT ★ 5 -
Agents make claims. Reelier writes receipts — record an agent's tool-call workflow once, replay it deterministically at 0 tokens, and diff runs to catch drift.
npm MIT ★ 3 -
A flight recorder for AI agents: records every step, ranks the step most likely at fault, then proves it by replaying the run with only that step repaired, at zero model calls. Top-1 0.60 on unseen failures against 0.25 baselines. MCP server, CI regression firewall, on PyPI.
MIT -
See what your Node backend really does on every request, catch what a change broke across every route, and prove fixes by replay. A local recorder, dashboard and MCP server, with a stop hook for coding agents.
npm MIT -
Browser automation and API testing MCP server on Playwright + Chrome DevTools Protocol. Numbered element refs and Set-of-Mark screenshots so AI agents target the right node; flow record/replay, element memory, captcha handoff, HTTP API test suite.
npm ISC -
@eddym06/custom-chrome-mcpgithub.com/eddym06/devtools-advance-mcpupdated this year
Cross-platform Chrome MCP with real-time request interception, response modification, network capture, HAR recording/replay, accessibility testing, anti-detection, and 91 tools. Includes smart workflows for common tasks. Works on Windows, Mac, and Linux.
npm MIT -
Cross-platform Chrome MCP that runs on a persistent clone of your real Chrome profile (session/logins carried over) with real-time request interception, response modification, network capture, HAR recording/replay, accessibility testing, stealth mode, con
npm MIT -
Compile one recorded browser-agent run into a deterministic BrowserOps procedure that replays with zero LLM tokens, plus a registry where procedures earn trust from verified runs, get demoted on site drift, and heal from the live page.
-
Aside-class browser agent — MCP · CLI · REPL. A real logged-in browser for AI agents: Keychain credential vault (secrets never in model context), enforced approval + read-only evidence mode, review evidence bundles, research/claim-drift, record→replay, and browsing-history recall.
MIT ★ 2 -
maponcegithub.com/shivampatel0714/maponceupdated this year
AI authors browser automations once (record or map); deterministic Playwright replays them at zero token cost - self-healing when the UI drifts, with business rules, credential housing, and an MCP server. LLMs at build time, auditable code at runtime.
MIT -
Realistic data for your Salesforce sandboxes: clone a record and its graph or a whole org, generate, anonymize and replay datasets, empty objects or an org, checkpoint an org and rewind it. A VS Code extension with production guardrails.
npm MIT ★ 2 -
Audit trail database for AI agents. Tamper-evident, checksum-backed decision logs with session replay and time-travel debugging to support EU AI Act Article 12 record-keeping. Drift detection, MCP, Python & TypeScript SDKs. Self-host or cloud.
npm AGPL-3.0 ★ 164 -
Automate Blender 3D modeling with AI via MCP — 93 tools for modeling, sculpting, architecture/MEP, materials, and 3D-print prep. Includes Blender Studio: a visual record/replay session editor with a multi-provider AI assistant (Claude, Gemini, OpenRouter) — drive it from n8n, Claude Code, or standalone.
pypi MIT ★ 86 -
neruva-mcpupdated this year
Neruva MCP -- proprietary memory + grounding substrate for AI assistants (Claude Desktop / Claude Code / any MCP client). Typed Records store, document knowledge graph with exact multi-hop traversal (ownership/relationship tracing), semantic + structural recall, deterministic corrections, provenance/citations, GDPR/Law-25 atomic forget, snapshot/replay audit. Thin client: all algorithms run server-side. For Neruva integration clients only -- NOT open source. Bring your own LLM.
pypi Proprietary -- (c) Clouthier Simulation Labs. For Neruva integration clients onl -
Bowline.Mcp — exposes the Bowline workbench itself as an MCP server, so AI agents (Claude Desktop, Cursor, Copilot) can drive discovery, invocation, recording, mocking, and replay through the Model Context Protocol. Sibling of Kuestenlogik.Bowline.Mock; the third row of the MCP surface area: Bowline as MCP server for its own operations (Kuestenlogik.Bowline.Protocol.Mcp covers the other two — Bowline as MCP client and as MCP adapter for fremd-protocols). Built on the official ModelContextProtocol C# SDK (1.2.x). Mount via `services.AddBowlineMcp() + endpoints.MapBowlineMcp(...)` for HTTP/SSE binding or run `bowline mcp serve` for stdio transport.
nuget apache-2.0 -
Bernsteinno update data
Deterministic, verifiable orchestration for CLI coding agents (Claude Code, Codex, Gemini CLI, Aider, Cursor and 40+ more), exposed over MCP. Bernstein runs orchestration tasks in isolated git worktrees, records a lineage and replay journal for every run, and can anchor task claims, progress updates and posted artifacts in an opt-in HMAC-chained audit log. Tools: 14 are advertised at the default `standard` tool tier. Set BERNSTEIN_MCP_TOOL_TIER to change the tier: `core` advertises 5, `standard` advertises 14 (default), `all` advertises 18. Transports: stdio (default), Server-Sent Events, and streamable HTTP. Install and run: uvx --from bernstein bernstein mcp or pip install bernstein && bernstein mcp HTTP transport: bernstein mcp --transport http Authentication: the stdio transport needs none. The HTTP transport accepts a static bearer token via BERNSTEIN_MCP_TOKEN or BERNSTEIN_AUTH_TOKEN, and binds to loopback only when neither is set. Documentation: https://bernstein.readthedocs.io Requires Python 3.12 or newer. Package `bernstein` on PyPI, version 3.9.0.
smithery -
Reproduce and record a browser bug once, verify the fix in seconds. A deterministic eye for coding agents — one MCP call instead of a model round trip per click.
MIT -
osier-devactive
Osier records what your coding agents do and tells you what's worth checking
npm MIT -
A free desktop CLI that automatically creates mocks and tests from watching your app run.
npm Apache-2.0 -
Reproducible game capture with scenarios, replayable sessions, OBS recording, and FFmpeg derivatives.
npm MIT -
maestrogithub.com/rexchengm/maestroupdated this year
Stronger hands for your AI — the RPA it can actually drive. Record a task once; AI optimizes, combines, and composes it into deterministic automations that control your real apps & browser. Pure-vision, local, MCP-native. The smarter the AI, the stronger it gets.
Apache-2.0 ★ 3 -
Voice-anchored causal evidence recorder for AI coding agents — talk, point, and get an evidence package your agent can act on. 语音锚定的因果证据记录仪。
npm MIT ★ 1