AVAILABLE CREDITS
—Prepaid account balanceLOCAL MVP
ZERODAY MODEL / INFERENCE INFRASTRUCTURE
Your models.
Your command.
One API for abliterated models. Create a key, choose your model, and keep every request and credit in view.
COMPLETED REQUESTS
—Across your workspaceTOKENS PROCESSED
—Reported by the providerYour next steps
From account to first requestTHE MODEL LIBRARY
Choose your engine.
Abliterated checkpoints through a familiar API. Only configured models can receive requests.
INFERENCE LAB / 01
Playground_
Build a conversation, tune the request, inspect the payload. Test model responses are fixed connectivity checks.
Configuration
SESSION ONLYConversation stays in this tab. Reloading clears it. Changing models starts a new conversation.
Inspect request JSON
Conversation
READY
[ Z / LAB ]
Start your next experiment.
Your conversation and model responses will appear here.
No requests in this session.
Session cost: $0.00
BUILD WITH ZERODAY
API keys
Keys belong to your account and use its credits. A new key is shown only once.
| Name | Key prefix | Created | Status |
|---|
EVERY REQUEST, ACCOUNTED FOR
Usage
Your latest 100 requests. Failed requests release their reserved credits.
| Model | Time | Input | Output | Cost | Status |
|---|
PREPAID, PAY AS YOU GO
Credits & billing
Prepaid inference. Clear balances. Every credit accounted for.
THIS MONTH / UTC
—Settled inference spendMONTHLY REQUESTS
—Completed requestsALL-TIME SPEND
—Across your accountAVAILABLE BALANCE
—Credits are reserved before a live request and settled against provider-reported tokens.
FUND YOUR WORKSPACE
Add API credits
Choose a credit pack and pay with a card or cryptocurrency. Credits arrive after the provider confirms payment.
Checking payment configuration…
Network fees may apply to crypto payments. Returning from checkout does not itself issue credits.
Payment history
Latest 100 payments · confirmed by provider| Created | Method | Amount | Status |
|---|
Credit activity
Latest 100 ledger entries · USD| Time | Movement | Amount | Reference |
|---|
A FAMILIAR INTERFACE
Your first API call.
Use an OpenAI-compatible client with your OpenForge key and the gateway URL below.
Python
Available endpoints
GET /v1/models
List enabled models for your account.
POST /v1/chat/completions
Text chat completions with usage and credit settlement.
Gateway limits
30 requests per minute per account. Two concurrent requests. Maximum 4,096 output tokens. Text and function tool messages.
Streaming uses buffered SSE: the provider completion is settled before chunks are delivered. Function tools are supported on models configured for tool calling.
YOUR ACCOUNT. YOUR TOOLS.
Connect your workflow.
Use your OpenForge account with OpenCode or any tool that accepts an OpenAI-compatible API.
1. Install & sign in
Install the downloadable CLI package. Browser approval creates a revocable CLI key stored in your OS credential vault.
2. Choose your model
The test model checks connectivity only. Coding requires a real model with function tool support.
Approve a CLI connection
Only approve a code shown by a login you started in your own terminal. The CLI receives a key that can spend your account credits.
OpenCode configuration
This configuration references OPENFORGE_API_KEY. It contains no secret. The CLI launcher supplies your key to the OpenCode process.
YOUR WORKSPACE IDENTITY
Account settings
Verify your email and manage account recovery.
Sign in to your account
Email verification status appears here.
Reset your password
Request a recovery link by email. Resetting your password signs out existing sessions and revokes all API keys.
Verify your email.
Confirm the verification link sent to your email address.
Set a new password.
Use at least 12 characters. Existing sessions and API keys will be revoked.