DEV Community

ideata.io
ideata.io

Posted on AI-assisted

I open-sourced Ideata: a self-hosted AEO tracker and AI content studio

Search is moving into the chat box.

People increasingly ask ChatGPT, Claude, Gemini, and Perplexity for recommendations instead of scrolling through ten blue links. I wanted to understand what those systems say about a brand, which sources they cite, and why competitors get recommended.

Most tools I found in this space were closed SaaS products. The methodology was hidden, the data lived on somebody else's infrastructure, and pricing was usually tied to seats, prompts, or reports.

So I built and open-sourced Ideata: a self-hosted AEO tracker with an AI content studio and cross-publishing workflow built in.

Ideata dashboard

What Ideata does

Track visibility across answer engines

Ideata runs real buyer prompts across 8+ answer engines, including ChatGPT, Claude, Gemini, Perplexity, DeepSeek, Grok, Yandex Alice/Neuro, and GigaChat.

For every run, it helps you inspect:

  • whether your brand was mentioned or cited
  • which competitors appeared instead
  • share of voice, position, and sentiment
  • the sources each model relied on
  • AI crawler activity and AI-referred traffic

Turn gaps into content

Tracking alone did not feel useful enough. Once Ideata finds a visibility gap, the built-in content studio can take an article through:

  1. brief and research
  2. outline
  3. draft
  4. quality checks and fact-checking
  5. publishing

The same article can then be adapted for Telegram, X, LinkedIn, Bluesky, Mastodon, VK, Threads, Zen, WordPress, or a custom REST endpoint.

The goal is a continuous loop: monitor → find gaps → create → distribute → measure again.

Why I made it open source

A visibility tool should not be another black box.

With Ideata:

  • your data stays on your infrastructure
  • you bring your own model provider keys
  • there is no phone-home behavior
  • the tracking and scoring logic is inspectable
  • you can modify the workflow for your own market or language

The project is licensed under AGPL-3.0 and the complete stack is available on GitHub.

Run it with Docker Compose

The stack uses Vue 3, NestJS, PostgreSQL, Redis, Prisma, and BullMQ. One Docker Compose command starts the dashboard, API, worker, database, and queue.

git clone https://github.com/Dallowin/ideata.git
cd ideata
cp .env.example .env
docker compose up -d --build
Enter fullscreen mode Exit fullscreen mode

Open http://localhost:8080, create the first owner account, and add an OpenRouter key or native provider keys under Settings → Providers.

You only need one model key to start. Native keys can be configured per engine when you need direct routing or grounded web search.

It is still early

Ideata is usable today, but there is plenty I want to improve: more answer engines, better geo-targeting, scheduled reports, and simpler deployment options.

I would especially appreciate feedback on:

  • which AI visibility metrics are actually useful
  • which answer engines or regional assistants are missing
  • whether the monitor-to-content workflow feels clear
  • what would make self-hosting easier

Issues and pull requests are welcome.

GitHub: github.com/Dallowin/ideata

If you work on SEO, AEO, content, or self-hosted software, I would love to hear how you would use it.

Top comments (0)