JSPM

  • ESM via JSPM
  • ES Module Entrypoint
  • Export Map
  • Keywords
  • License
  • Repository URL
  • TypeScript Types
  • README
  • Created
  • Published
  • Downloads 17
  • Score
    100M100P100Q52141F
  • License MIT

A local-first, token-efficient LLM chat interface

Package Exports

  • opengauge
  • opengauge/core

Readme

OpenGauge

A local-first, token-efficient LLM chat interface.

If you just want to use it, you only need one command:

npx opengauge

What this repo does

OpenGauge runs a local web chat app with:

  • Token optimization (compression + deduplication + checkpoints)
  • Context retrieval from conversation history (RAG-style)
  • Multiple providers: Anthropic, OpenAI, Gemini, Ollama
  • Local storage in SQLite at ~/.opengauge/opengauge.db
npx opengauge

You do not need to install OpenGauge globally first. npx will download the package (if not already cached) and run it. On first run, it may take a little longer while it fetches the package.

This starts a local server and opens the app in your browser.

Default URL:

http://localhost:3000

If port 3000 is busy:

npx opengauge --port 3001

First-time setup

When the app opens, configure your provider in the UI wizard, or create:

~/.opengauge/config.yml

Example:

providers:
  anthropic:
    api_key: YOUR_API_KEY
    default_model: claude-opus-4-6

defaults:
  provider: anthropic

Developer setup (from source)

git clone https://github.com/applytorque/opengauge.git
cd opengauge
npm install
npm run build
npm start

Useful commands

npm run build        # Compile TypeScript + copy UI assets
npm start            # Run CLI entry locally
npm pack --dry-run   # Preview npm package contents

Competitors and positioning

OpenGauge sits between chat UIs and full LLM observability suites.

PromptOps / Observability tools

Examples: PromptLayer, Helicone, Langfuse, Humanloop, Arize Phoenix

  • Great at: traces, eval pipelines, team dashboards, observability depth.
  • OpenGauge advantage: local-first workflow, in-chat prompt improvement, duplicate-risk + token-efficiency feedback in one loop.

Multi-model chat interfaces

Examples: LibreChat, Open WebUI, Chatbot UI, AnythingLLM

  • Great at: broad chat UX and plugin ecosystems.
  • OpenGauge advantage: prompt quality optimization is first-class (Improve + analytics), not only chat.

IDE coding assistants

Examples: Continue and other IDE-native AI assistants

  • Great at: deep coding workflows inside editors.
  • OpenGauge advantage: model-agnostic PromptOps for any user (product, ops, research, support), not just coding.

Cloud model platforms

Examples: Azure AI Studio, Vertex AI, Bedrock consoles

  • Great at: enterprise governance and managed cloud workflows.
  • OpenGauge advantage: fast setup, no cloud lock-in, supports local Ollama and cloud providers together.

Why choose OpenGauge

  • Improve prompts before send (optional Auto Improve mode)
  • Measure quality, duplicates, and token efficiency after send
  • Keep data local with SQLite and run quickly with npx opengauge

Positioning line:

OpenGauge is PromptOps in the loop: improve prompt quality before send, measure impact after send, and reduce token waste continuously.

License

MIT