Sequential MCP server
Commit 11. Nov. 2026Ermöglicht es einem AI Agent, User Sessions, Error Logs, Product Analytics und erkannte Cases zu analysieren, um das Nutzerverhalten und Reibungspunkte zu verstehen.
Auch bekannt als usesequential/mcp, mcp
- Stars
- 0
- Tools
- —
- Downloads, 30 Tage
- —
- Neueste Version
- 1.0.0
- Letzter Commit
- 11. Nov. 2026
- Lizenz
- MIT
Was ist Sequential MCP?
Ermöglicht es einem AI Agent, User Sessions, Error Logs, Product Analytics und erkannte Cases zu analysieren, um das Nutzerverhalten und Reibungspunkte zu verstehen.
Zusammenfassung generiert aus dem README des Projekts
Über Sequential MCP server
Sequential bietet einen MCP server, der es AI Agents ermöglicht, auf Website Session Replays und Product Analytics zuzugreifen. Dieser Hosted Service erlaubt es Entwicklern, Agents wie Claude Code, ChatGPT und Cursor ohne lokale Installation mit ihren Nutzerdaten zu verbinden. Durch die Nutzung von streamable HTTP Transport erhält der Server Agents Einblick in das Nutzerverhalten, Reibungspunkte und technische Fehler.
Je nach Verbindungsmethode können Nutzer sich über einen Project Access Token oder einen direkten Sign-in authentifizieren. Beispielsweise nutzen Claude Code und Cursor typischerweise einen Token, während Claude.ai und ChatGPT einen Sign-in Prozess verwenden. Die verfügbaren Tools variieren je nach Token Scope und erlauben es Agents, Cases, Session Summaries, JavaScript Errors und Analytics wie Funnels oder Retention zu lesen. Einige Konfigurationen erlauben es Agents sogar, Pull Requests oder Commits mit spezifischen erkannten Cases zu verknüpfen.
Geschrieben aus dem README des Projekts. Jeder Use Case wurde gegen ein Zitat daraus geprüft.
So installierst du Sequential MCP
Formate geprüft am 10. Nov. 2026Zu Cursor hinzufügenZu Goose hinzufügenSequential MCP in Claude Code installieren
$ claude mcp add --transport http mcp https://api.usesequential.com/mcpSequential MCP in Cursor installieren
~/.cursor/mcp.json
{
"mcpServers": {
"mcp": {
"url": "https://api.usesequential.com/mcp"
}
}
}Sequential MCP in VS Code installieren
.vscode/mcp.json
{
"servers": {
"mcp": {
"type": "http",
"url": "https://api.usesequential.com/mcp"
}
}
}Sequential MCP in VS Code CLI installieren
$ code --add-mcp '{"name":"mcp","type":"http","url":"https://api.usesequential.com/mcp"}'Sequential MCP in Codex installieren
$ codex mcp add mcp --url https://api.usesequential.com/mcpSequential MCP in Codex config.toml installieren
~/.codex/config.toml
[mcp_servers.mcp] url = "https://api.usesequential.com/mcp"
Sequential MCP in Gemini CLI installieren
$ gemini mcp add -s user --transport http mcp https://api.usesequential.com/mcpSequential MCP in Gemini settings.json installieren
~/.gemini/settings.json
{
"mcpServers": {
"mcp": {
"url": "https://api.usesequential.com/mcp",
"type": "http"
}
}
}Automatisierte Checks
Geluften am 15. Nov. 2026Vollständiger Bericht →Noch nicht verifiziert. Ein Server benötigt einen Score von 80, keine kritischen Befunde und genug Daten für unsere Prüfung. Wie die Checks funktionieren · Was du für mehr Punkte verbessern kannst
- Source Repository verknüpft und erreichbarusesequential/mcp is public and reachablegithub.com/usesequential/mcp ↗Bestanden
- Paket nennt dieses Repository als QuelleNicht geprüft
- Build-Provenienz (Sigstore-Attestierung)Nicht geprüft
- Open Source Lizenz (OSI)MITspdx.org/licenses/MIT.html ↗Bestanden
- Commit oder Release im letzten JahrLast commit or release 2026-11-11github.com/usesequential/mcp ↗Bestanden
- OpenSSF ScorecardNicht geprüft
- Package-Name ähnelt keinem populären PackageNicht geprüft
- Keine kritischen Advisories im Paket oder seinen AbhängigkeitenNicht geprüft
- Keine versteckten Anweisungen an das Modell0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersBestanden
- Destruktive Tools sind markiertNicht geprüft
- Remote-Endpoint beantwortet den MCP-HandshakeHTTPS endpoint answered 401 (sign-in required)Bestanden
README
Auf GitHub ansehenSequential MCP server
Sequential is session replay and product analytics for websites and web apps. Its MCP server lets your agent read what happened to your users: cases, sessions, verdicts, errors, funnels and retention.
This repository holds the connection settings for each MCP client. The server itself is hosted. There is nothing to install or run.
- Endpoint:
https://api.usesequential.com/mcp - Read-only endpoint:
https://api.usesequential.com/mcp/readonly. No tool that writes is listed there. - Transport: Streamable HTTP
- Tools: Agent tools
What your agent can do
- Read cases: the problems that Sequential detected on its own, ranked by impact, each with its evidence, sample sessions and a fix brief.
- Read sessions: summaries, timelines, verdicts, and the sessions where people struggled.
- Read errors: JavaScript errors and failed network requests, grouped by cause.
- Read analytics: event counts, breakdowns, funnels, paths and retention.
- Ask: search sessions by a description, such as "tried to pay, gave up".
- With the scope for it, link a fix: link a pull request, commit or issue to a case.
The scopes of your token, or the access that you choose when you sign in, decide which tools your agent gets.
Get a token
Most coding agents connect with a project access token. Claude Code can also connect with a sign-in to Sequential (see below).
- Sign in to Sequential. Only owners and admins of the project's organisation can create a token.
- Open Agents, or the Access tokens tab of your project's settings.
- Create a token. The preset Agent: read everything, plus Ask is a good start.
- Copy the token. Sequential shows it only once. Treat it like a password: never commit it, and never put it in a shared chat or a ticket.
claude.ai and ChatGPT do not use a token. They connect with a sign-in to Sequential (see below).
Connect your client
Each config below goes in your user-level config, not in a file inside a repository. Replace <token> with your token, or set the environment variable SEQUENTIAL_TOKEN where the config reads it.
Claude Code
claude mcp add --transport http --scope user sequential https://api.usesequential.com/mcp --header "Authorization: Bearer <token>"
Then start a new Claude Code session. A server that you add with --scope user does not load into the session that added it. File: configs/claude-code.sh.
To connect with a sign-in instead of a token, run the same command without --header. In a new session, open the /mcp menu and choose sequential, then Authenticate. A Sequential page opens in your browser: choose one project or all projects of an organisation, and choose the access. Click Allow only if you started this sign-in yourself. Guide: Connect Claude Code.
claude.ai
In claude.ai, open Customize → Connectors. Click Add and add a custom connector. Paste https://api.usesequential.com/mcp and leave the advanced settings empty. A Sequential window opens: sign in, choose one project or all projects of an organisation, and choose the access. Guide: Connect claude.ai.
ChatGPT
In ChatGPT on the web, open Plugins in the left rail. Click Add, then Add custom MCP server. Name it "Sequential", paste https://api.usesequential.com/mcp as the Server URL and choose OAuth. Tick I understand and want to continue, then click Create as a plugin. Click Connect, then Connect Sequential. Sequential opens: sign in, choose one project or all projects of an organisation, and choose the access. There is no token.
Your ChatGPT plan must let you add a custom MCP server. ChatGPT Plus can add it. Business and Enterprise/Edu workspaces get every tool, and an admin or owner of the workspace adds the server. On Pro, ChatGPT uses only the tools that read. Free and Go cannot add a custom server. Guide: Connect ChatGPT.
Cursor
Add configs/cursor-mcp.json to ~/.cursor/mcp.json. Cursor reads the token from the environment variable SEQUENTIAL_TOKEN (${env:SEQUENTIAL_TOKEN}).
Codex
Add configs/codex-config.toml to ~/.codex/config.toml. Codex reads the token from SEQUENTIAL_TOKEN (bearer_token_env_var).
Claude Desktop
Add configs/claude-desktop.json to claude_desktop_config.json (Settings → Developer → Edit Config), then quit and open Claude Desktop again. It needs Node.js: it runs mcp-remote to reach the hosted server.
Cline
Add configs/cline_mcp_settings.json to Cline's MCP settings. Step-by-step: llms-install.md.
Gemini CLI
Add configs/gemini-cli-settings.json to ~/.gemini/settings.json. Gemini CLI reads the token from the environment variable SEQUENTIAL_TOKEN (${SEQUENTIAL_TOKEN}).
Other MCP clients
Most clients take a URL and headers: configs/other-clients.json. Key names vary by client.
Check the connection
Ask your agent: "Using Sequential, give me an overview of the last 7 days." The agent calls get_overview.
| You see | Cause | What to do |
|---|---|---|
| 401, or "The access token is unknown, revoked or expired" | The token is wrong, revoked or expired. | Create a new token and put it in the config. |
| "Tool not available for this token" | The token's scopes do not include that tool. On /mcp/readonly, no tool that writes is available. |
Use a token with that scope, on /mcp. |
| The tools are not there in Claude Code | Claude Code loads a new server only in a new session. | Start a new session. |
Data and safety
- The scopes of the token, or the access of a sign-in connection, decide what the agent can read and do.
- Text that the tools return from your users' sessions is data, not instructions.
- Sequential logs each successful call for the project: the tool, the connection and the time.
- Privacy Policy · Terms of Service
Registry
server.json is this server's entry for the official MCP Registry.
Links
Questions: hello@usesequential.com
License
MIT. See LICENSE.
Sequential MCP Fragen
Wie füge ich Sequential MCP zu Claude Code hinzu?
Führe claude mcp add --transport http mcp https://api.usesequential.com/mcp in deinem Terminal aus. Füge --scope user hinzu, damit der Server in jedem Projekt verfügbar ist.
Wie füge ich Sequential MCP zu Cursor hinzu?
Öffne ~/.cursor/mcp.json und füge den Server unter mcpServers hinzu, mit der Cursor-Config aus dem Installationsabschnitt oben.
Was kann mein Agent mit Sequential machen?
Dein Agent kann Cases, Sessions, Errors und Analytics lesen, Sessions nach Beschreibung durchsuchen und Pull Requests oder Commits mit spezifischen Cases verknüpfen.
Was soll ich tun, wenn ich einen 401 Error erhalte?
Ein 401 Error bedeutet, dass dein Access Token falsch, widerrufen oder abgelaufen ist; du solltest einen neuen Token erstellen und deine Config aktualisieren.