Now in public beta

Build AI pipelines.
Ship without infrastructure.

Chain models, prompts and logic into multi-step workflows.
Trigger via API, cron, or one click. No servers needed.

Start building free → View documentation

weekly-blog-generator Running

🔍 Research claude-sonnet-4

💡 Pick Topic gpt-4o

✍️ Write Draft claude-sonnet-4

✅ Quality Review claude-sonnet-4

Works with Anthropic OpenAI Meta Llama Mistral

HOW IT WORKS

From YAML to production
in three steps

Define your pipeline as code. We handle execution, retries, scheduling, and delivery.

Define

Write your pipeline in YAML or JSON. Chain models, set variables, configure retries and output formats.

Run

Trigger manually, via API, cron, or webhook. Steps execute in sequence with real-time streaming.

Deliver

Output goes where you need it — webhooks, email, API response. Track cost and tokens per step.

PIPELINE AS CODE

Everything you need to
orchestrate AI workflows

pipeline.yaml

name: weekly-blog-generator
version: 1

variables:
  language: "fr"
  tone: "direct, pragmatic"

steps:
  - id: research
    model: claude-sonnet-4
    output_format: json
    retry:
      max_attempts: 2

  - id: write_draft
    model: claude-sonnet-4
    output_format: markdown

Model-agnostic

Mix Claude, GPT-4o, Llama and more in a single pipeline. Each step picks the best model for the job.

Variable interpolation

Reference previous outputs, user inputs, secrets, and pipeline metadata with Handlebars-like syntax.

Retries & conditions

Built-in retry with backoff. Conditional branching lets you loop or skip steps based on output quality.

Cron scheduling

Schedule pipelines with cron expressions. Timezone-aware, with distributed locking for reliability.

FEATURES

Built for production

Everything beyond the basics to run AI workflows at scale.

Real-time streaming

Watch each step execute live via SSE. See token counts, costs, and outputs as they stream in.

Encrypted secrets

Store API keys and tokens with AES-256-GCM encryption. Reference them safely in prompts via {{env.xxx}}.

YAML import/export

Define pipelines in YAML or JSON. Import, export, and version control your workflows alongside your code.

Cost tracking

Per-step token counts and costs. Get estimates before running. No surprise bills.

API & webhooks

Trigger pipelines programmatically. Unique webhook URLs per pipeline. Full REST API with scoped keys.

Version history

Every edit creates an immutable snapshot. Roll back to any version. Compare definitions side by side.

PRICING

Start free, scale as you grow

100 free credits to start. No credit card required.

Free

€0 /month

100 credits included

3 pipelines
5 steps per pipeline
10 runs/day
Community models
Run history

Get started

Pro

€49 /month

8,000 credits included

Unlimited pipelines
20 steps per pipeline
500 runs/day
All models
Cron scheduling
Webhooks & API access
Priority execution

Start free trial

Enterprise

Custom

Custom credit allocation

Everything in Pro
SSO / SAML
Dedicated workers
SLA guarantee
Custom model endpoints
Priority support

Contact sales

Build AI pipelines.Ship without infrastructure.

From YAML to productionin three steps