# BenchGen

## Docs

- [Action Asset Library](https://benchgen.com/docs/agentspace/action-asset-library.md): Reuse preconfigured Actions instead of rebuilding from scratch.
- [Agent Concepts](https://benchgen.com/docs/agentspace/agent-concepts.md): Understand what an agent is, how it works, and when to use one.
- [Agent Connections](https://benchgen.com/docs/agentspace/agent-connections.md): Securely integrate your Agentspace agents with external platforms and services.
- [Agent-to-Agent (A2A) Connection](https://benchgen.com/docs/agentspace/agent-to-agent-integration-guide.md): Step-by-step guide to connecting your Agentspace agent using the SSE protocol.
- [Connect Custom Coder Agent](https://benchgen.com/docs/agentspace/connect-custom-coder-agent.md): Connect your custom UBOS Agentspace coder agent to VS Code using the Cline extension.
- [How to Create an Action](https://benchgen.com/docs/agentspace/how-to-create-action.md): Step-by-step guide to adding a new Action to a Topic in Agentspace.
- [How to Create an Agent](https://benchgen.com/docs/agentspace/how-to-create-agent.md): Step-by-step guide to creating an agent from scratch or using a template in Agentspace.
- [How to Create a Topic](https://benchgen.com/docs/agentspace/how-to-create-topic.md): Step-by-step guide to adding a new Topic to your agent in Agentspace.
- [What Is a Topic](https://benchgen.com/docs/agentspace/topic-concepts.md): Learn how Topics define what your agent can do, when it engages, and how it behaves.
- [Using AI to Generate a Topic](https://benchgen.com/docs/agentspace/using-ai-to-generate-topic.md): Bootstrap a Topic definition from a natural language prompt using the Generate with AI feature.
- [Using Topic Templates](https://benchgen.com/docs/agentspace/using-topic-templates.md): Accelerate consistent Topic creation using pre-built classification, scope, and instruction patterns.
- [Overview](https://benchgen.com/docs/agentspace/welcome.md): Your dedicated area for managing, building, and exploring AI agents on the UBOS platform.
- [What Is an Action](https://benchgen.com/docs/agentspace/what-is-an-action.md): Learn how Actions are the executable units that let your agent do real work.
- [Create Plant](https://benchgen.com/docs/api-reference/endpoint/create.md): Creates a new plant in the store
- [Delete Plant](https://benchgen.com/docs/api-reference/endpoint/delete.md): Deletes a single plant based on the ID supplied
- [Get Plants](https://benchgen.com/docs/api-reference/endpoint/get.md): Returns all plants from the system that the user has access to
- [New Plant](https://benchgen.com/docs/api-reference/endpoint/webhook.md): Information about a new plant added to the store
- [Introduction](https://benchgen.com/docs/api-reference/introduction.md): Example section for showcasing API endpoints
- [Add a Model](https://benchgen.com/docs/eval/add-a-model.md): Register a new model in Eval by uploading a model archive or importing one from HuggingFace.
- [Bundle Structure](https://benchgen.com/docs/eval/bundle-structure.md): The files inside a custom environment .zip bundle and what each one does.
- [Create a Custom Environment](https://benchgen.com/docs/eval/create-environment.md): How to build and upload your own evaluation environment to BenchGen.
- [Evaluate an Inference Model](https://benchgen.com/docs/eval/evaluate-a-running-model.md): Benchmark a live inference endpoint against an Eval environment, monitor the run in real time, and read the scored results.
- [Export Datasets → Train](https://benchgen.com/docs/eval/export-datasets.md): Export failing benchmark cases as a labeled dataset to kick off a fine-tuning run in Train.
- [Monitor Model Usage](https://benchgen.com/docs/eval/monitor-model-usage.md): Track requests, tokens, latency, and spend for a deployed model, and inspect individual request logs.
- [Overview](https://benchgen.com/docs/eval/overview.md): What Eval does, how environments work, and what it hands off to Train and Agents.
- [Read Results](https://benchgen.com/docs/eval/read-results.md): Understand the benchmark results report and what the metrics mean.
- [Run a Benchmark](https://benchgen.com/docs/eval/run-a-benchmark.md): Step-by-step guide to running a model against a benchmark in Eval.
- [Deploy an Inference Model](https://benchgen.com/docs/eval/run-an-inference-model.md): Spin up a live, OpenAI-compatible inference endpoint for a model in Eval, monitor its deployment, and verify it is serving requests.
- [YAML Reference](https://benchgen.com/docs/eval/yaml-reference.md): Every field in competition.yaml for a BenchGen custom environment bundle.
- [Introduction](https://benchgen.com/docs/index.md): The infrastructure for self-improving agents. BenchGen is a Synthetic Data Factory that creates a digital twin of your business.
- [Agent Learning Loop](https://benchgen.com/docs/self-improvement-agent.md): How BenchGen's six-stage Simulate → Act → Capture → Measure → Generate → Improve cycle builds self-improving agents from your enterprise data.
- [Add a Dataset](https://benchgen.com/docs/train/add-a-dataset.md): Register a training dataset by importing it from HuggingFace or uploading your own file.
- [Fine-tune a Model](https://benchgen.com/docs/train/fine-tune-a-model.md): Configure and launch a LoRA fine-tuning run, then monitor it to completion.
- [Merge & Save a Model](https://benchgen.com/docs/train/merge-lora-adapter.md): Merge a trained LoRA adapter into its base model and push it to BenchGen so you can run and evaluate it.
- [Overview](https://benchgen.com/docs/train/overview.md): What Train does, when to use it, and what it hands off to Eval and Agents.
- [Run Inference](https://benchgen.com/docs/train/run-inference.md): Serve a model you trained as a live, OpenAI-compatible endpoint and send it requests.

## OpenAPI Specs

- [openapi](https://benchgen.com/docs/api-reference/openapi.json)