OpenGEO

Open source AI search visibility — measure, act, rewrite, re-check

OpenGEO tracks whether AI search mentions and cites your brand across ChatGPT, Perplexity, and Google AI Overviews. Then it turns gaps into an action brief and citable content blocks, and re-checks the same prompts after you ship.

See the gap. Ship the block. Prove the citation.

Why OpenGEO?

• Prompt-first. GEO starts with questions people ask AI.
• Mentions + citations + share of voice. Know who gets named and which URLs get sourced.
• Action Brief. Every gap becomes a concrete next step with evidence.
• Citable Blocks. Definition, comparison, FAQ, and sourced snippets ready to publish.
• Re-check loop. Run the same prompts again and see what actually moved.
• MCP & Agent Skills. Your agent runs the workflow; you review in the UI.
• Pay for what you run. Demo mode ships locally. Wire live providers when ready. Fork it. Self-host it.

Quick start

Requires Node 22 (see .nvmrc). Node 24 may fail to install native SQLite bindings without local C++ build tools.

pnpm install
cp .env.example .env
pnpm db:seed
pnpm dev

API: http://127.0.0.1:8787
Web (Vite): pnpm dev:web → http://127.0.0.1:5173

Default API key from .env.example: dev-opengeo-key.

Docker

docker compose up --build

Main workflows

1. Create a project (brand, domain, competitors)
2. Seed / curate prompts
3. (Optional) Audit public GEO readiness (llms.txt, robots, sitemap)
4. Run visibility across engines
5. Inspect citation gaps + share of voice
6. Generate an Action Brief
7. Generate Citable Blocks and publish them on your site
8. Re-check the same prompts and read the diff
9. Export a Markdown / JSON report

MCP

pnpm mcp

Example MCP client config:

{
  "mcpServers": {
    "opengeo": {
      "command": "pnpm",
      "args": ["mcp"],
      "cwd": "/absolute/path/to/OpenGEO",
      "env": {
        "DATABASE_PATH": "./data/opengeo.db",
        "VISIBILITY_PROVIDER": "demo"
      }
    }
  }
}

Tools

Tool	Purpose
whoami	Identity + engines
list_projects / create_project / get_project	Projects
list_prompts / upsert_prompts / suggest_prompts	Prompt library
run_prompt_check / run_visibility_batch / get_job_status	Visibility runs
get_visibility_snapshot / get_share_of_voice / get_citation_gaps	Results
generate_geo_brief	Action Brief
generate_citable_blocks	Publishable blocks
run_recheck / get_recheck_diff	Post-publish proof
export_visibility_report	Markdown / JSON export
audit_geo_readiness	Public readiness checks

Agent Skills

Install skills into your agent skills directory from .agents/skills/:

• /geo-project-setup
• /prompt-research
• /geo-readiness
• /visibility-check
• /citation-gap
• /geo-brief
• /citable-blocks
• /geo-recheck
• /geo-coach

Costs

Visibility runs can cost money once live providers are connected. Local demo mode is free and deterministic so you can validate the full loop without API keys.

Billing unit for live mode: prompt × engine.

License

MIT