LAUNCHING SOON — $20 USD / month. Subscriptions are not open yet.
Coding plan For OMP & Hermes agents

Stop counting tokens.

One flat-price agent plan that removes token anxiety. Unlimited reasonable access to Fabryka AI models — let your OMP or Hermes agent read, plan and refactor without watching a meter.

$ 20 /month
flat
Explore the $20 plan → $20 USD/month · Fair use applies
omp — ~/repo/your-project
$ omp
› Refactor the billing module and add tests.
· model: qwen3.8-27b
· read → plan → edit → test
· your agent, your workspace
model access: one monthly plan
›
WORKFLOW DEMOIllustration
Extra cost$0.00
Plan$20 flat
01 / PRICEOne number. $20.No tiers to decode, no credits that expire.
02 / ACCESSQwen. Bielik. Fabryka.Including Qwen 3.8 27B. One familiar API.
03 / AGENTSBuilt for OMP & Hermes.Your agent, connected to the Fabryka API.
What “reasonable” means

Built for a developer, not a data centre.

Made for one person’s everyday coding. Shared capacity and model limits apply. Resale, shared keys and bulk generation are outside the plan. Final usage limits will be published before subscriptions open.

✓Covered Everyday coding-agent sessionsRefactors, tests and multi-file plansQwen 3.8 27B and included modelsPersonal and professional projects
×Not covered Reselling or sharing your keyServing production customer trafficAutomated bulk generation

Running in three steps.

01 Get the plan The $20 monthly plan is launching soon. Paid access instructions will be available before checkout opens.
plan.fabryka.ai
02 Connect your agent Configure OMP with a custom OpenAI-compatible provider and your Fabryka API key.
https://fabryka.ai/v1
03 Choose your model Start with Qwen 3.8 27B. Run your coding workflow within the plan’s fair-use limits.
qwen3.8-27b

Questions.

Is it really unlimited?

The plan is designed for reasonably unlimited everyday coding by one person. Shared capacity, model context limits and fair-use restrictions apply. Final usage limits will be published before checkout opens.

Which models are included?

The launch plan is centered on Qwen 3.8 27B and Bielik 11B v3. The included model list will be confirmed before subscriptions open; features and availability vary by model.

Does it work with OMP and Hermes?

Fabryka has an OpenAI-compatible API. OMP and Hermes can use a custom provider; the exact settings depend on your agent version. Follow the Fabryka setup guide and use qwen3.8-27b as the model ID.

Will long agent loops get throttled?

This is shared model access, not dedicated compute. Busy periods and heavy sessions may require waiting or retrying. Fair-use and per-request limits apply; there is no per-token overage billing on the plan.

How do billing and cancellation work?

Stripe will handle the $20 USD monthly subscription and recurring payments. Any applicable taxes are shown before payment. Customers will be able to manage billing and cancel future renewals through the Stripe customer portal.

Can I subscribe today?

Not yet. Checkout is closed while we finish billing and subscriber access. You can try the existing public Fabryka API now; it is separate from the forthcoming paid plan.

Let the agent run. The bill won’t.

Try Fabryka while we prepare → Launching soon · Checkout is not open