Shape the crew
Give each agent a role, a goal and a model — or lift a ready crew from 330+ templates and adjust it to taste.
Work in a steady current
Let your busywork flow downstream.
Flowdy is a no-code studio where AI agents pick work up, pass it along and set it down finished. Describe the crew once; from then on, the current carries it — and one token balance meters every run.
Top up once, run anytime330+ ready crewsNo subscriptions
From still to flowing
Four gentle steps, then the water takes over
Lay the channel
Link tasks so output glides into the next step. Add the services each agent touches along the way.
Open the gate
Trigger by hand, by schedule or by API call, then watch each step stream its work in real time.
Read the meter
Each run settles into the ledger as an exact whole-token debit with your balance after it. Top up whenever it suits you.
Watch the current
One crew, source to delivery
Follow a morning-digest crew through a complete cycle — from the first agent to the ledger line it leaves behind.
0+
crews ready to put in the water
0+
services the current can reach
0/7
runs on schedule while you sleep
No
monthly fees — tokens only
The channel
Six ways Flowdy keeps work moving
Nothing here demands your attention. Each piece is built to be set once and left running.
Drawn, not scripted
Sketch the path work should take on a canvas — agents, tasks, hand-offs. The drawing is the pipeline; there is no second config to maintain.
Meets your tools midstream
120+ connectors let a crew read RSS, call any HTTP endpoint, and deliver into Slack or Telegram without a line of glue code.
Your models, your keys
Plug in Anthropic, OpenAI, Gemini or xAI credentials and assign each agent the model suited to its step. Keys are AES-encrypted at rest.
Runs while you rest
Schedule crews or let webhooks wake them. Morning digests, overnight research, hourly monitors — the current does not keep office hours.
A balance, not a bill
Every run draws whole tokens from one balance, and every movement lands in an append-only ledger showing the balance after it.
Failure costs nothing
Tokens are reserved before a run and released the moment one fails. Unfinished work is never charged — the ledger proves it.
Tokens
Fill the reservoir, spend as it flows
There is exactly one thing to buy here: power tokens. They sit on your balance until a crew draws on them — no renewal date, no plan, no seats.
Maker — €25
2,100 tokens+100 bonus tokens≈ 5 typical runs
- Hierarchical orchestration
- Public API + 3 keys
- Crew memory & templates
Bigger top-ups simply lower the price per token. Unused tokens never expire.
Quiet questions
Things people ask before they wade in
No. You top up a token balance and crews draw it down as they run. Nothing renews, nothing expires, and there is no plan to downgrade — the balance simply sits until you use it.
Whatever the ledger says, to the whole token. Model usage and connector work convert into power tokens, and each run writes one append-only row with the debit and your balance after it.
No. Crews are drawn on a visual canvas, and 330+ templates cover common pipelines end to end. The REST API exists for the moment you want to trigger crews from your own software — it is optional.
Yes. Add keys for Anthropic, OpenAI, Gemini, xAI and other OpenAI-compatible providers, then route each agent to the model you prefer. Stored keys are AES-encrypted.
The reserve taken at the start of the run is released. Failures never bill — you only ever pay for work that completed.
Connector credentials are AES-encrypted at rest, and API keys are stored as hashes — the plaintext appears once, at creation, and never again.
Put one stream of work on autopilot tonight.
Create an account, pick a template, and let your first crew run before you close the laptop.
Claim starter tokens