Nyckel
Free tierThe AI platform for reliable predictions — build custom ML models without the PhD
Free tier available·All audiences·Powered by GPT (OpenAI), Claude (Anthropic), Gemini (Google), CLIP, DINO·API available
Key strengths
Automatic model selection and retrainingActive learning for fast model improvementSupports images, text, and structured dataSOC2 certified and HIPAA compliantNo ML expertise requiredConfidence scores and fast inference
Free tier + paid plans · from $149 USD/mo
Facts verified against nyckel.com on
Recent changes
- Plans updated: Free tier at $0/mo (100 invokes, up to 200 samples, up to 5 functions); Starter at $149/mo (15,000 invokes, up to 30k samples, up to 20 functions); Business at $599/mo (300,000 invokes, up to 600k samples, up to 100 functions); Enterprise with custom pricing. Overage: $0.005/invoke on Free/Starter, $0.001/invoke on Business.
- Models updated: GPT (OpenAI), Claude (Anthropic), Gemini (Google), CLIP, DINO
Get alerts when Nyckel changes
Free. We'll email you when pricing, the free tier, API access or open-source status changes.
No ratings yet
Nyckel provides a REST API for creating and invoking "functions" — the platform's term for a trained ML model endpoint. Key integration points include:
- Function creation: Define a function with a name, input type (image, text, or structured data), and output type (classification, detection, prediction).
- Sample ingestion: POST labeled samples to train the model; active learning surfaces high-value samples for annotation.
- Invocation: Send inference requests to the function endpoint and receive predictions with confidence scores.
- Automatic retraining: The platform selects and retrains the best-fit model (from GPT/OpenAI, Claude/Anthropic, Gemini/Google, CLIP, DINO) as new samples are added — no manual pipeline management.
- Plan limits: Functions per account are capped at 5 (Free), 20 (Starter), and 100 (Business). Invoke overages are billed at $0.005/invoke on Free/Starter and $0.001/invoke on Business.
- Compliance: SOC2 certified and HIPAA compliant; suitable for regulated data environments.
- Open source: An open-source component is available on GitHub (3 stars).
Alternatives
See all alternatives