Bring your own API keys. Any LLM, any provider.

Build AI agents visually. Deploy as REST APIs.

Design multi-step AI pipelines on a drag-and-drop canvas. LLM calls, tool use, conditional branching, loops, parallel execution, and human review checkpoints.

Every agent becomes a live API endpoint. One POST request, that's it.

Start building free Sign in
Your agent, live as an API
curl -X POST https://your-app.com/api/agents/AGENT_ID/run \
  -H "X-AgentHub-Key: ahk_xxxxxxxxxx" \
  -d '{"message": "Summarise this article..."}'

// Response
{ "output": "...", "tokens": 342, "status": "completed" }

Works with any LLM. You bring your own API keys.

OpenAIGeminiClaudeGroqLlamaMistralOllamaAny OpenAI-compatible API

How it works

Three steps from idea to running API

1

Design on the canvas

Drag nodes onto the canvas and connect them. Each node does one thing: call an LLM, run a tool, check a condition, or wait for human approval.

2

Pick any LLM

Configure your own API keys for OpenAI, Gemini, Claude, Groq, Mistral, or any OpenAI-compatible endpoint including self-hosted Ollama.

3

Save and call it

Every agent is instantly available as a REST API. One POST request and your agent runs, tools fire, LLMs respond, results stream back.

8 node types. Infinite combinations.

Every building block you need to model any AI workflow

LLM
Call any language model with a system prompt
Tool
Web search, HTTP calls, code execution, databases
Condition
Binary yes/no routing evaluated by LLM
Switch
Multi-way routing for 3+ branches
Loop
Repeat until exit condition is met
Fork / Join
Parallel execution, merge results
HITL
Pause for human approval before continuing
Clarify
Ask user a follow-up question mid-run

Everything you need

Built for engineers who want to ship AI fast

Visual DAG Builder
Build complex multi-step pipelines by connecting nodes. No code required for the flow.
Instant REST API
Every agent auto-gets a live POST endpoint. Call it from Python, JavaScript, curl, or any HTTP client.
Any LLM Provider
Bring your own API keys. Mix providers across nodes: GPT-4 for reasoning, Gemini for speed.
Conditional Routing
Condition and Switch nodes let the agent pick its own path based on what it finds.
Loops & Parallel
Loop nodes refine until quality passes. Fork/Join nodes run multiple branches simultaneously.
Human-in-the-Loop
HITL nodes pause the run for human review before continuing. Full audit trail included.
Guardrails
Block bad input and filter unsafe output at the node level. Per-node, not per-agent.
Runs & Analytics
Every execution is logged with tokens, latency, cost, and full trace. Filter and export.

Ready to build your first agent?

No credit card. No infra to manage. Sign up, build, and call your agent in under 5 minutes.

Get started free
Free to start
Your own API keys
Instant REST API
No vendor lock-in

Get in touch

Have a question, feature request, or want to report a bug?
We'd love to hear from you.