Build a crew of AI agents. Pay by the power, never by the month.
Compose role-based agents, hand them tasks, and run them as a team — sequential or hierarchical. Top up power tokens once; SaasHub meters every run down to the token.
From idea to running crew in four steps.
No Python, no infrastructure. Everything happens in your browser; the heavy lifting runs in our queue.
Build agents
Give each agent a role, a goal and a backstory. Pick its model and the tools it's allowed to use.
Define tasks
Describe the work and the expected output. Chain tasks with dependencies so context flows downstream.
Assemble a crew
Drop agents and tasks into a crew. Choose a sequential pipeline or a manager-led hierarchy.
Run & pay per token
Hit run. Watch the live trace, and pay only for the tokens each step actually consumed.
Connect your apps and AI agents — no code, no glue.
A guided walkthrough: pick a trigger, wire services and agents together with branching logic, then watch it run and bill by the token. Press play, or jump to any chapter.
Pick a trigger
no codeConnect apps & AI agents
drag to connectRun it live
streaming tracePay per token
reserve releasedWatch every run, token by token.
A live trace of each step — which agent acted, which model it called, what it cost. No black boxes, no surprise bills.
Market brief crew
runningEverything you need to orchestrate agents — and nothing you don't.
Roles, goals & backstory
Shape each agent's behavior with a role, a clear goal and a backstory — the same mental model as CrewAI, no code required.
Sequential & hierarchical
Run tasks in order, or let a manager agent plan and delegate. Switch the process type with a single toggle.
Built-in tools
Web search, HTTP fetch, calculator and a sandboxed code runner — attach them per agent and meter each call.
Live step trace
Every reasoning step, tool call and delegation streams into a timeline you can audit while it runs.
Transparent token ledger
An immutable record of every top-up, debit and refund — reserved up front, released when a run comes in under budget.
Public API & keys
Trigger crews programmatically with scoped API keys. Same metering, same ledger — now from your own stack.
Steal a crew. Make it yours.
Battle-tested crew templates wired and ready. Clone one into your workspace, swap the models, and run.
Market brief crew
Researches a topic, extracts drivers & risks, and ships an executive brief — reviewed for accuracy.
Blog factory
Outlines, drafts, and edits an SEO-aware article from a single keyword, with a fact-check pass.
Bug triage manager
A manager reads the issue, delegates a repro and a fix proposal, then summarizes for the team.
Lead enrichment
Takes a company name, gathers firmographics, scores fit, and drafts a tailored first-touch email.
Buy power tokens. Spend them whenever.
No seats, no monthly fee, no expiry. Top up a balance and every run draws from it at cost-plus. The more you buy, the bigger the bonus.
- ~2 typical crew runs
- All 6 model providers
- Sequential process
- Built-in tools & actions
- ~6 typical crew runs
- Hierarchical orchestration
- Public API + 3 keys
- Crew memory & templates
- ~12 typical crew runs
- Unlimited API keys
- Priority queue lane
- Low-balance alerts
Every new account starts with 1,000 free tokens. Tokens never expire and refund automatically when a run finishes under its reserve.
Questions, answered.
How is a token priced, and what does a run cost?
One power token is a flat $0.0002. A run draws tokens for every model call and tool use across the crew — a typical 3–4 agent crew lands around 400–700 tokens. You see the exact charge per step in the live trace.
Do tokens expire or auto-renew?
Never. There are no subscriptions and no monthly resets — you top up a balance once and it stays until you spend it. If a run finishes under the amount it reserved, the difference is refunded to your balance automatically.
Which models and providers can I use?
OpenAI, Anthropic and Google Gemini are supported out of the box, and you can mix models within a single crew — a cheap model for research, a stronger one for writing. Bring your own provider keys, or run on ours and pay in tokens.
What's the difference between sequential and hierarchical?
Sequential runs agents in a fixed order, passing each result downstream. Hierarchical puts a manager agent in charge — it plans the work, delegates sub-tasks to the right agents, and aggregates their output. Toggle between them with one click.
Can I trigger crews from my own app?
Yes. Every crew has a public API endpoint and scoped API keys. The same metering and token ledger apply, so a programmatic run is billed exactly like one launched from the dashboard.
Do I need to write any code?
No. Agents, tasks and crews are all built visually in the browser. Code is optional — reach for the API and the sandboxed code runner when you want them, not because you have to.
Spin up your first crew in minutes.
Build the agents, drop them into a crew, and let SaasHub run the work. You only pay for the tokens it burns.