Simple, transparent pricing.
Start free and scale as you grow. No hidden fees, no surprises.
Free
$0 /mo
Get started instantly
No credit card required
- Basic models (8 models)
- 16K context window
- 100 requests/day
- API keys
- Community support
Recommended Pro
Coming Soon
For professionals and teams
Notify me →Everything in Free, plus:
- All 50+ models
- Full context windows (up to 10M)
- Higher rate limits
- Priority inference
- Streaming + function calling
- Email support
Max
Coming Soon
For power users and enterprises
Notify me →Everything in Pro, plus:
- 10x higher rate limits
- Priority queue access
- Dedicated capacity
- Custom model hosting
- SLA guarantee
- Dedicated support
Compare Plans
Feature breakdown.
Feature
Free
$0/mo
Pro
Coming soon
Max
Coming soon
Models
8 basic
50+ all
50+ all
Context window
16K
Up to 10M
Up to 10M
Rate limit
100/day
10,000/day
100,000/day
Streaming
Function calling
Priority inference
Custom models
Support
Community
Email
Dedicated
SLA
FAQ
Common questions.
The free tier gives you immediate access to 8 open-source models with up to 100 requests per day. No credit card is required — just sign up, get an API key, and start building. Free tier models support up to 16K context windows and full streaming responses.
Pro and Max plans are currently in development and will launch soon. Sign up for notifications on the pricing cards above to be the first to know when they become available. We’re focused on getting the free tier rock-solid first before expanding paid offerings.
Yes. idyl.inference implements the OpenAI chat completions API specification. Any tool, SDK, or library that works with the OpenAI API will work with idyl — just change the base URL. This includes the official OpenAI Python and Node SDKs, LangChain, LlamaIndex, and coding tools like Cursor, Claude Code, and Aider.
The free tier includes 8 carefully selected open-source models spanning chat, code, and reasoning capabilities: Qwen 3.5 9B, Qwen 3.5 4B, Devstral 24B, Codestral 22B, Gemma 3 27B, Phi-4 14B, Mistral Small 24B, and DeepSeek R1 8B. These represent the best models in their size classes.
Absolutely. You can upgrade or downgrade your plan at any time. When upgrading, you’ll get immediate access to the new tier’s features. When downgrading, your current billing cycle continues until the end of the period, then switches to the new plan. Your API keys and configurations stay the same.
Yes. For teams that need dedicated infrastructure, custom model hosting, volume discounts, or specific SLA requirements, we offer tailored enterprise plans. Contact our sales team to discuss your needs and get a custom quote.