anomalyco / anomalyco/models.dev
Deepgram provider is missing from the catalog
Nobody has claimed this yet.
- Dominant language
- TypeScript
- Stars
- 6.9k
- Forks
- 1.7k
- Avg merge
- 3h 21m
- Merged PRs (30d)
- 873
Description
Problem
https://models.dev/api.json has no deepgram provider, even though Deepgram has an official AI SDK provider (@ai-sdk/deepgram) and publishes a machine-readable model catalog.
AI SDK provider docs: https://ai-sdk.dev/providers/ai-sdk-providers/deepgram
Deepgram model metadata docs: https://developers.deepgram.com/guides/fundamentals/model-metadata
Available source data
GET https://api.deepgram.com/v1/models requires no API key and returns separate stt and tts groups. Each entry includes the callable canonical_name, architecture, languages, version, and whether applicable, batch/streaming flags. On 2026-08-27 it returned 143 unique canonical IDs, including:
- STT:
nova-3-general,nova-3-medical, Nova 2 variants, Whisper variants - TTS: Aura and Aura 2 voice IDs such as
aura-2-thalia-en
Deepgram also documents the model families and callable IDs:
- STT: https://developers.deepgram.com/docs/models-languages-overview
- TTS: https://developers.deepgram.com/docs/tts-models-languages-overview
Schema consideration
These are audio models, so mandatory language-model token limits and text-token cost fields do not fit every entry. Modalities do fit (audio -> text for STT and text -> audio for TTS), but the provider may expose a broader schema gap around non-token billing and non-LLM limits.
RubyLLM currently has to discover this catalog directly from Deepgram because models.dev cannot supply it.
Contributor guide
No contributing guide indexed for this repository
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating the existing provider catalog ingestion entry points and schema definitions, then inspect GET https://api.deepgram.com/v1/models and the linked Deepgram model documentation. Add the Deepgram provider with its STT and TTS canonical IDs, modalities, and available metadata, while addressing the noted non-token billing and limit fields. Done means the catalog represents the returned models without requiring an API key.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- typescript
- Domain
- ai, backend-api-design, data
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Active
- Clarity
- Mostly clear
- Newbie friendliness
- 52/100