flux

Features

Everything Flux does.

From the first node on the canvas to the production trace — Flux owns every step of the agent lifecycle so your team ships, not babysits.

Visual canvas

Drag nodes for inputs, models, tools, conditionals and outputs. Switch to code view any time — Flux YAML round-trips cleanly with the visual.

  • Live debug — replay any node with last input
  • YAML in repo, visual on canvas, both round-trip
  • Real-time collaboration — multi-user editing

Multi-model router

Route by task, latency or cost. Fall back across providers automatically. Bring your own keys (OpenAI, Anthropic, Mistral, Google) or use the metered Flux router.

  • BYO keys for every major model provider
  • Auto-fallback on rate limits and outages
  • Cost-aware routing per task type

200+ tool integrations

Slack, Notion, Linear, Salesforce, Postgres, Stripe, GitHub. Add custom tools as JSON-schema-typed functions — agents call them like native primitives.

  • 200 prebuilt integrations, OAuth-based auth
  • Custom tools via JSON-schema or OpenAPI spec
  • Per-team marketplace for shared tool packages

Eval suite

Golden datasets, regression checks, prompt diffing. Catch broken agents before they hit production — and roll back to the last green version with one click.

  • Golden datasets per agent flow
  • Regression checks on every PR
  • One-click rollback to last green deploy

One-click deploy

Vercel, Cloudflare, Modal, AWS Lambda or your own VPC. Flux generates the runtime, you ship the trigger — webhooks, cron, manual or event-driven.

  • Vercel, Cloudflare, Modal, Lambda one-click
  • VPC / on-prem deployment on Enterprise
  • Triggers: webhook, cron, queue, event

Observability built in

Trace every agent run end-to-end. Token usage, latency, tool calls, errors. Replay failed runs in one click and patch the flow before the next deploy.

  • Per-run trace timeline (spans, prompts, tool calls)
  • Token + cost + latency rollups per agent
  • One-click run replay with frozen inputs

Ship your first agent before lunch.

Free for individuals. Fork a template, swap the model, deploy to Vercel — the whole loop in less time than a build pipeline.