IgnitionAI

Ownership of code, models and data

  • The source code is yours

    All the code we write during an engagement is handed over to your company at delivery. No limited usage license, no royalty, no conditional reversibility clause. You can modify it, redistribute it internally, or even shut it down — it's your asset.

  • Your data never leaves

    During the engagement, your data stays on your infrastructure or on a cloud you control. We systematically sign an NDA before any access, and a DPA if personal-data processing is involved. No copy, no snapshot, no leak to our internal systems.

  • Fine-tuned models belong to you

    If we train a model on your data (fine-tuning, specialised embeddings), the resulting weights are your intellectual property. They are deployed at your site, not ours. We keep no trace of them after delivery.

  • Open-source building blocks stay open-source

    We're happy to build on open-source components (LangGraph, vLLM, Qdrant, etc.) because it's the best way to avoid lock-in. Their original license applies — usually permissive (MIT, Apache 2.0). No proprietary IgnitionAI stack imposed.

  • And after the engagement?

    You can take everything back in-house, bring in another vendor, or keep working with us. No exit penalty, no mandatory maintenance contract. We stay available if you want to call us back, that's all.

Ownership of code, models and data — IgnitionAI