TypeSafe AI provider for the Backdrop CMS AI module.
Adds TypeSafe AI's decision models to the providers the ai module can route
to, using the /v1/systemone endpoint at https://api.typesafe.ai. These
models return typed judgments with calibrated probabilities instead of
generated text, which suits fast classification and moderation checks.
| Operation | Supported | Notes |
|---|---|---|
| Decisions | Yes | boolean, choice and score questions via decide(); choice questions take options, score questions optional levels (2-10, low to high). |
| Moderation | Yes | Implemented as a single boolean decision; score is the probability of a violation. |
| Chat | No | |
| Completions | No | |
| Tool calling | No | |
| Vision | No | |
| Embeddings | No | |
| Image generation | No | |
| Speech-to-text | No | |
| Text-to-speech | No |
The model list is fetched from TypeSafe AI's /v1/models endpoint, so new
models appear without a module update. When no model is chosen, the first
listed model is used. See https://docs.typesafe.ai/models.
The decision endpoint and request timeout are stored in
ai_provider_typesafeai.settings (endpoint, timeout).
- Install this module using the official Backdrop CMS instructions.
- Create an authentication key with the Key module holding your TypeSafe AI API key (https://console.typesafe.ai/keys).
- Enable and configure the provider at
admin/config/ai/settings, and select a default decision or moderation model as needed.
Bugs and feature requests should be reported in the Issue Queue.
-
Created for Backdrop CMS by Justin Keiser.
-
Developed with AI assistance.
This project is GPL v2 software. See the LICENSE.txt file in this directory for complete text.