OctoPulse

A zero-backend GitHub maintainer dashboard. Triage what needs you and what's rotting across many repos from a single page — no server, no data collection, no vendor lock-in.

• Action Required Inbox — review requests, assignments, blocked PRs, and unanswered external issues
• Repository Pulse Grid — health-at-a-glance for every tracked repo (green / amber / red)
• Stale Watch — open threads ranked by days-since-last-activity
• One-click merge / close — confirm modal, REST mutation, refresh
• LLM-powered thread summarizer + intent classifier — bring-your-own provider (OpenAI / Groq / OpenRouter / Gemini / any OpenAI-compatible endpoint, including local llama.cpp / Ollama)
• Read-only or read & write — pick a PAT lane at sign-in; write actions are hidden when read-only
• Demo mode — drive the full UI with no token

Everything runs in your browser. Your PAT and LLM key live in localStorage and are never transmitted anywhere except directly to the APIs they belong to.

---

Get started

Pick the path that fits you. All three give you the same app.

1. Use the hosted version (zero setup)

https://octopulse.kpsn.dev

Open it, paste a PAT (or try demo mode), done. The whole app is static — the host serves bytes, never sees your token.

2. Self-host with Docker (one command)

docker run -d --name octopulse -p 8080:80 ghcr.io/kingpin/octopulse:latest

Then open http://localhost:8080. Multi-arch image (amd64 + arm64), nginx-served, ~63 MB. Updates with docker pull.

Prefer compose? Drop this in docker-compose.yml:

services:
  octopulse:
    image: ghcr.io/kingpin/octopulse:latest
    ports: ["8080:80"]
    restart: unless-stopped

3. Self-host from source (static files)

git clone https://github.com/KingPin/OctoPulse.git
cd OctoPulse
npm install
npm run build        # outputs static dist/

Serve dist/ from any static host. Common options:

• GitHub Pages — Settings → Pages → Source: GitHub Actions. .github/workflows/deploy.yml ships with this repo and builds on every push to main. Custom domain: edit public/CNAME. Project page (no domain): delete public/CNAME and set base: '/<repo>/' in vite.config.ts.
• Cloudflare Pages — build command npm run build, output dist, no env vars.
• Vercel — framework preset Vite, output dist, no functions.
• Just on your laptop — cd dist && python3 -m http.server 8000.

For local dev with hot reload: npm run dev → http://localhost:5173.

---

GitHub Personal Access Token

OctoPulse needs a PAT to call the GitHub API directly from the browser. (Device Flow would be nicer but github.com/login/device/code blocks browser CORS.) Choose one of:

Classic PAT — quickest. Visit https://github.com/settings/tokens/new?scopes=repo,read:org,read:user&description=OctoPulse. The scopes repo, read:org, read:user are pre-filled. Generate, copy, paste into OctoPulse.

Fine-grained PAT (recommended for least-privilege) — Visit https://github.com/settings/personal-access-tokens/new. Grant access to the repos you want to track and the following permissions:

• Issues — Read & write (needed to close issues)
• Pull requests — Read & write (needed to merge PRs)
• Contents — Read
• Metadata — Read
• Members — Read (for org repos)

The token stays in your browser's localStorage under the key octopulse:githubToken. Use the Sign out button to clear it.

---

LLM provider setup

Open the settings panel (gear icon, top right) → AI Assistant section. The summarizer + intent classifier are gated on a configured provider; the dashboard itself works without one.

Provider	Where to get a key	Default model
OpenAI	https://platform.openai.com/api-keys	gpt-4o-mini
Groq	https://console.groq.com/keys	llama-3.3-70b-versatile
OpenRouter	https://openrouter.ai/keys	openrouter/auto
Gemini	https://aistudio.google.com/apikey	gemini-2.0-flash
Custom (OpenAI-compatible)	local — llama.cpp, LM Studio, Ollama, etc.	llama3 (default base URL http://localhost:11434/v1)

Click Test after pasting your key. Defaults rot fast — feel free to override the model. Keys persist in localStorage under octopulse:llmConfig.

Ollama users: start Ollama with OLLAMA_ORIGINS="*" ollama serve (or your specific origin) so the browser can call it.

---

Privacy & security

• Your PAT and LLM API key live in localStorage only. They never leave your browser except to the APIs they authenticate to (api.github.com, your configured LLM provider).
• No telemetry, no analytics, no third-party scripts.
• The whole app is a single static bundle; you can read the network tab and verify.
• The hosted version at octopulse.kpsn.dev is the exact build from main — same bytes you'd get from npm run build or the Docker image.

If you don't want OctoPulse to keep a token at all, use demo mode.

---

Stack

Vite 8 · React 19 · TypeScript strict · Tailwind v4 · Lucide React · Vitest

GitHub data: GraphQL primary (/graphql) with REST fallbacks for the few endpoints GraphQL doesn't cover (merge PR, close issue, fetch issue body, fetch comments).

Refresh strategy: load + manual button + on tab-focus if the cache is older than 5 minutes.

---

License

Apache License 2.0 — see LICENSE.