nimbus

nimbus

glama.json already carries one, so this may prefill

Category
访问服务器

README

<div align="center">

Nimbus

Cross-service incident context in under 100 ms. Consent-gated automation. Your credentials never leave the machine.

A local-first, HITL-gated AI agent over your dev tools — it builds a private index of your work across 80+ services and answers questions and runs multi-step workflows entirely on your machine.

CI License: AGPL-3.0 Docs Discussions

Install · Docs · Watch the 15-second cast · Architecture

</div>


[!NOTE] Nimbus is a headless Gateway + CLI (plus a VS Code extension and a browser web clipper). It runs on your machine and talks to ~80 cloud services through first-party MCP connectors. Nothing about your data — the index, your credentials, the audit log — leaves your box.

What it does

Three things, in one query:

  • Incident response — PagerDuty alert → deploy → commit → author, correlated locally.
  • CVE / code exposure — indexed code search across every connected repo, with no fan-out network calls.
  • Data lineage — Tableau → Looker → dbt → Airflow → the renamed column.

Three load-bearing words

  • local — the SQLite index, the Vault, and the audit log all live on your machine. The cloud is a connector, not the source of truth. Telemetry is opt-in and off by default ([telemetry] enabled = false).
  • consent-gated — every destructive or outbound action is intercepted by a human-in-the-loop gate before it runs. It lives in the executor, not the prompt, so it can't be jailbroken away.
  • MCP — Nimbus speaks the Model Context Protocol in both directions. As an MCP client it drives every connector as an MCP server (tools), and hosts any third-party server you register with nimbus connector add --mcp. As an MCP server, nimbus mcp-server --stdio exposes your local index to editor AIs through six read-only tools. The engine never calls a cloud API directly.

Quickstart

1. Install (per-user, no admin/sudo required):

<details open> <summary><b>macOS / Linux</b></summary>

# Linux only — credentials live in the OS keystore, and the Gateway will not
# start without it:  sudo apt install libsecret-tools   (Debian/Ubuntu)
#                    sudo dnf install libsecret         (Fedora/RHEL)
# macOS only — the keychain must be UNLOCKED. Nimbus never shows an
# authorization dialog (a background service could not answer one), so on a
# locked keychain it fails immediately and tells you what to run. Over SSH or
# in CI, give it its own keychain:  security create-keychain -p "" nimbus.keychain
#                                   security default-keychain -s nimbus.keychain
#                                   security unlock-keychain  -p "" nimbus.keychain
curl -fsSL https://github.com/nimbus-agent/Nimbus/releases/latest/download/install.sh -o /tmp/nimbus-install.sh
# inspect it first if you like:  less /tmp/nimbus-install.sh
bash /tmp/nimbus-install.sh
nimbus --version

</details>

<details> <summary><b>Windows (PowerShell, no admin)</b></summary>

irm https://github.com/nimbus-agent/Nimbus/releases/latest/download/install.ps1 | Invoke-Expression
# open a new PowerShell window:
nimbus --version

</details>

Every release artefact is GPG-signed (key 5A20457CCD8B53FFAA945240886ADA6B487CAB6E) with a SHA-256 manifest and build-provenance attestations — see Verify your download. Homebrew and Scoop taps are also available (see the install guide).

2. Index a repo you already have — no account, no token, no API key:

cd ~/code/your-project
nimbus init

nimbus init adds the repo to nimbus.toml with code indexing on, starts the gateway, and indexes it. It appends to your config — it never rewrites it, so your comments and existing settings survive (and it keeps a nimbus.toml.bak).

3. Trace a line's provenance — who wrote it, the PR, the ticket, the incident it responded to, and what breaks downstream:

nimbus why src/auth.ts:42

nimbus init prints a real file:line from your own repo to try first. This works with no credentials and no LLM configured.

Optional: add an LLM

Indexing, nimbus why, and the agent briefs all work with no LLM configured — briefs render deterministically. An LLM buys you two things: nimbus ask (natural-language queries), and prose synthesis that rewrites those briefs into more readable narrative.

It does not have to be a cloud one. Point Nimbus at a local model and nothing — not even prompts — leaves the machine:

# ~/.config/nimbus/nimbus.toml
[llm]
prefer_local = true
local_model  = "llama3.1"     # served by Ollama on http://127.0.0.1:11434

See Local & air-gapped LLM setup.

Optional: connect a cloud service

To correlate across GitHub, Jira, PagerDuty, Slack and ~80 others, add a connector. The fastest path is a token-based one like GitHub:

nimbus connector auth github --token <your_PAT>
nimbus connector sync github
nimbus ask "what PRs did I open in the last 7 days?"

(OAuth services — Google Drive, Gmail, Slack, … — use nimbus connector add <service>, which opens your browser. See Connect a service.)

How it works

80+ cloud services ─▶ first-party MCP connectors ─▶ local SQLite index (+ embeddings)
                                                          │
                              your question ─▶ engine ─▶ HITL consent gate ─▶ action
                                                          │
                                       CLI · VS Code · web clipper · (desktop, coming)

A headless Bun Gateway maintains the private index and runs the agent; clients talk to it only over local JSON-RPC IPC. Credentials live in the OS keystore (DPAPI / Keychain / libsecret) — never in logs, config, or IPC. Full design: Architecture.

Connectors

80+ first-party MCP connectors across Google, Microsoft, GitHub/GitLab, Slack, Jira, Notion, plus observability, CI/CD, security/quality, feature-flags, GitOps, data/BI, deploy, finance, and support tools. Browse the full roster in the connector docs; building your own is covered in CONTRIBUTING.

Security & trust

  • Human-in-the-loop consent on every outbound/destructive action (structural, not a prompt).
  • No plaintext credentials — OS Vault only.
  • Signed, reproducible releases — GPG manifests, SBOM (CycloneDX), build-provenance attestations.
  • Audited — see the security model & disclosure policy.

Found a vulnerability? See SECURITY.md.

Community & contributing

License

Dual-licensed by design: AGPL-3.0 for the gateway, CLI, and MCP connectors; MIT for the separately-published @nimbus-dev/sdk and @nimbus-dev/client npm packages so extensions and integrations stay unencumbered.

推荐服务器

Baidu Map

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。

官方
精选
JavaScript
Playwright MCP Server

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。

官方
精选
TypeScript
Magic Component Platform (MCP)

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。

官方
精选
本地
TypeScript
Audiense Insights MCP Server

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。

官方
精选
本地
TypeScript
VeyraX

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。

官方
精选
本地
graphlit-mcp-server

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。

官方
精选
TypeScript
Kagi MCP Server

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。

官方
精选
Python
e2b-mcp-server

e2b-mcp-server

使用 MCP 通过 e2b 运行代码。

官方
精选
Neon MCP Server

Neon MCP Server

用于与 Neon 管理 API 和数据库交互的 MCP 服务器

官方
精选
Exa MCP Server

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。

官方
精选