seq-mcp
A Model Context Protocol server that provides AI agents with controlled read access to Datalust Seq instances for log analysis and monitoring. It enables agents to search events, execute data queries, and retrieve information about signals, dashboards, and alerts.
README
mcp-seq-otel
Standalone MCP server that gives AI agents controlled API access to a user-owned Datalust Seq instance.
What This Service Assumes
- You already run/manage your own Seq instance.
- This service only needs connection settings.
Required configuration:
SEQ_URLSEQ_API_KEY
URL Support
SEQ_URL accepts either:
- Host URL (recommended; service will append
/api):http://localhost:10150https://seq.example.com
- Full API base URL (also supported):
http://localhost:10150/apihttps://seq.example.com/api
Available MCP Tools
seq_connection_test: validates Seq connectivity and API reachability.seq_starter_help: lists the compact starter alias tools.seq_starter_overview: quick health, user, diagnostics, signals, and workspace summary.seq_starter_events_search: common event search by filter/signal/time range.seq_starter_event_by_id: fetch one event by id.seq_starter_data_query: run aqquery usingGETorPOST.seq_starter_signals_list/seq_starter_signal_by_id: signal discovery helpers.seq_starter_dashboards_list: list dashboards.seq_starter_alerts_list: list alerts.seq_starter_events_stream: bounded live-tail style stream call.seq_api_catalog: returns the full official Seq route/verb/permission catalog.seq_api_live_links: discovers livename -> routelinks from your Seq instance.seq_api_request: generic verb/path invoker for any Seq API route.seq_<verb>_<route>: auto-generated tool per official route+verb (from docs).
Scope note:
seq_starter_*tools are focused on common read workflows.seq_api_requestandseq_<verb>_<route>expose the broader HTTP API surface from the Seq endpoint catalog, including non-GETroutes.
Seq API Key Permissions
Use least privilege based on the exact tools/workflows your MCP client will call.
Authoritative Datalust references:
- API keys and permission model: https://docs.datalust.co/docs/api-keys
- HTTP API usage guide: https://docs.datalust.co/docs/using-the-http-api
- Server endpoint + permission table: https://docs.datalust.co/docs/server-http-api
Recommended permission profiles:
- Read-focused starter usage (
seq_starter_*,seq_connection_test, read-only queries): enableRead; disableIngest,Write,Project,Organization,System. - Full route-surface usage (
seq_api_requestandseq_<verb>_<route>): required permissions depend on the specific route+verb; checkseq_api_catalogor the official endpoint table before granting.
Permission guidance for this project:
| Permission | Needed now | Why |
|---|---|---|
Read |
Yes | Required by starter query/retrieval workflows. |
Ingest |
Usually No | Needed only when calling ingestion routes such as ingest/* or api/events/raw when API-key-for-writing is required. |
Write |
Maybe | Needed for write routes (for example signals, dashboards, alerts, permalinks, SQL queries). |
Project |
Maybe | Needed for project-scoped administration and some settings/index routes. |
Organization |
Maybe | Needed for organization/user-management routes. |
System |
Maybe | Needed for system administration routes (for example apps, feeds, backups, updates). |
Starter endpoint mapping in current implementation:
GET /health: public endpoint.GET /api/events/resources: public endpoint.GET /api/events:Readpermission demand.GET /api/events/{id}:Readpermission demand.GET /api/data:Readpermission demand.
API Key Header
The server authenticates to Seq using the X-Seq-ApiKey header.
Resilience and Permission Handling
Tool failures are returned as structured MCP error responses (isError: true) instead of crashing the server process.
Current graceful handling includes:
401 Unauthorized: returns guidance to verifySEQ_API_KEYandSEQ_URL.403 Forbidden: returns a permission-denied response with route-derived permission hints when available.- Network/timeout failures: returns connectivity diagnostics for AI clients.
Local Run (Node)
Node.js requirement: LTS (see .nvmrc).
npm install
npm run build
Windows PowerShell:
$env:SEQ_URL = "http://localhost:10150"
$env:SEQ_API_KEY = "your-key"
node dist/index.js
Docker Standalone Run
Docker Hub documentation assets in this repo:
- Auto-sync source for Hub long description:
.dockerhub-readme.md - Main project reference:
README.md
Build image:
docker build -t mcp/seq-otel:latest .
Run via Docker Compose (recommended for MCP stdio):
PowerShell:
./scripts/run-mcp-compose.ps1 -SeqUrl "http://localhost:10150" -SeqApiKey "<YOUR_SEQ_API_KEY>" -ImageTag latest -Build
Bash:
./scripts/run-mcp-compose.sh --seq-url "http://localhost:10150" --seq-api-key "<YOUR_SEQ_API_KEY>" --image-tag latest --build
Compose run behavior:
- Value precedence is: explicit script args -> existing environment variables ->
.env. - If the target
.envis missing, scripts create a generic template.envautomatically. - When that generated generic template is in use, ensure
SEQ_URLandSEQ_API_KEYare resolved via script args, existing environment variables, or a populated.envfile. MCP_IMAGE_TAGprecedence is: explicit-ImageTag/--image-tag-> existing environment ->.env->latest.- Scripts check image availability through compose service metadata and auto-build if the selected tag is not available.
- By default, no container name is specified, so Docker auto-generates a random name.
- You can override container name with
-ContainerName(PowerShell) or--container-name(Bash). - For security, prefer setting
SEQ_API_KEYvia environment variables or.envinstead of CLI args.
Or use helper scripts:
PowerShell:
./scripts/build-image.ps1
Bash:
./scripts/build-image.sh
Build and push to registry (pullable by other Docker hosts):
PowerShell:
./scripts/build-image.ps1 -Registry ghcr.io/mclifeleader -Tag v0.2.0 -Push
Bash:
./scripts/build-image.sh --registry ghcr.io/mclifeleader --tag v0.2.0 --push
Build and export tar (loadable with docker load):
PowerShell:
./scripts/build-image.ps1 -SaveTar ./mcp-seq-otel.tar
Bash:
./scripts/build-image.sh --save-tar ./mcp-seq-otel.tar
Run against local Seq:
docker run --rm -i \
-e SEQ_URL=http://host.docker.internal:10150 \
-e SEQ_API_KEY=your-key \
mcp/seq-otel:latest
Run against FQDN Seq:
docker run --rm -i \
-e SEQ_URL=https://seq.example.com \
-e SEQ_API_KEY=your-key \
mcp/seq-otel:latest
Run with Podman:
podman build -t mcp/seq-otel:latest .
podman run --rm -i \
-e SEQ_URL=https://seq.example.com \
-e SEQ_API_KEY=your-key \
mcp/seq-otel:latest
The container startup contract requires both variables to be present:
SEQ_URLSEQ_API_KEY
If either is missing, the container exits immediately with a clear startup error.
Copy/Paste MCP Config (Codex and VS Code)
Assumes the image already exists (mcp/seq-otel:latest).
{
"mcpServers": {
"seq-otel": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"--rm",
"-i",
"-e",
"SEQ_URL=http://host.docker.internal:10150",
"-e",
"SEQ_API_KEY=<YOUR_SEQ_API_KEY>",
"mcp/seq-otel:latest"
]
}
}
}
Docker MCP Catalog Compatibility
This repo includes a catalog-ready server definition in:
catalog/server.yamlcatalog/tools.jsonassets/seq-otel-icon.svg
catalog/server.yaml declares:
SEQ_URLas a required user parameter (mapped viaconfig.env)SEQ_API_KEYas a required secret (mapped viaconfig.secrets)
These files follow the Docker MCP registry server format, so they can be used in
catalog generation/import workflows (for example via docker/mcp-registry tools).
Submission checklist for Docker MCP Registry:
docs/docker-mcp-registry-submission.md
MCP Client Integration Example
Use a command-based MCP client entry that launches the container with stdin/stdout attached.
{
"mcpServers": {
"seq-otel": {
"type": "stdio",
"command": "docker",
"args": [
"run",
"--rm",
"-i",
"-e",
"SEQ_URL=https://seq.example.com",
"-e",
"SEQ_API_KEY=<YOUR_SEQ_API_KEY>",
"mcp/seq-otel:latest"
]
}
}
}
Security Notes
- Do not commit keys.
- Use least-privilege Seq API keys.
- Prefer HTTPS for non-local Seq instances.
API Mapping Artifact
- Full generated API map:
docs/api-map.md - Maintenance/update playbook:
docs/seq-mcp-maintenance.md - Project skill for update workflows:
.github/skills/seq-mcp-maintainer/SKILL.md
Status
Core standalone server scaffold is implemented and buildable.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。