Get started
BenchGen AI Agent
Chat with BenchGen’s built-in agent to get oriented fast.
Agents Overview
Build and deploy AI agents with topics, actions, and integrations.
Agents
Connect an OpenClaw Agent
Wire up an existing OpenClaw agent to BenchGen.
Benchmark an OpenClaw Agent
Run it against an Eval environment and read the trace.
Datasets from an OpenClaw Agent
Turn real traces into training data.
How to Create an Agent
Build your own agent from scratch with the Agent Builder.
Eval
Overview
What Eval does and where to go next in this section.
Run a Benchmark
Score a model against an environment, on the platform, HuggingFace, or an external API.
Add a Model
Register a model by uploading an archive or importing from HuggingFace.
Create a Custom Environment
Package and upload your own evaluation environment.
Train
Overview
What Train does and where to go next in this section.
Add a Dataset
Register a training dataset from HuggingFace, a file, or agent traces.
Upload a Model
Bring your own base model into Train.
Fine-tune a Model
Configure and launch a LoRA fine-tuning run.
Merge & Save a Model
Download the adapter, merge it, or push it to BenchGen.
Run Inference
Serve a trained model as a live, OpenAI-compatible endpoint.
Inference
Overview
What Inference covers and where to go next in this section.
Inference API Token
Issue a key from a model’s page for direct API usage.
BenchGen Router Lite
An OpenRouter-style endpoint that picks the model for you.
Deploy
Helm
Deploy the BenchGen platform with Helm.
DT Edge Platform
Deploy the digital-twin edge platform.
Helm Values
Every configurable value in the Helm chart.
Guides & API
Router Models Guide
The dataset, training, and full benchmark story behind BenchGen Router Lite.
API Reference
The public, unauthenticated part of the BenchGen API.