policy transform active execution plan live

AI EXECUTION CONTROL · USAGEFLOW VIBE

Change what your AI does
without shipping code.

Control models, tools, and spend at runtime.
Per customer. Per policy. Every request.

Read docs
TypeScript @usageflow/vibe Python usageflow-vibe Go usageflow-go-middleware/v2 OpenAI OpenAI • Anthropic Anthropic

Free: 1 workflow with live policy • No credit card

Scroll

Intent → transform → why

Every request is evaluated in real time
and adapted to the right model, tools, and limits.

Live example · customer-support · identity cust_acme · no redeploy

Requested

  • Modelopus-4
  • ReasoningHigh
  • BrowserOn
  • Tokens8k
UsageFlow Awaiting

Executed

  • —
01 intent Recording intent usage.chat() · opus · reasoning=high · browser · 8k
02 policy Checking policy cost_ladder · identity cust_acme at $98 / $100 cap
03 transform Rewriting plan haiku · reasoning=low · browser off · 2k tokens
04 why Explaining change budget exceeded · cost_ladder · no redeploy

Developer sends intent — Console policy has not run yet.

WHAT VIBE IS FOR

Make every request fit the moment.

Adapt models, tools, and limits to each customer, context, and policy.

FREE CUSTOMER · budget 82%

GPT-5 GPT-5 mini
requested executed ↓ 64% cost

PRO CUSTOMER · campaign active

GPT-5 mini GPT-5
requested executed ↑ premium experience

ENTERPRISE · sensitive data

Web search tool Blocked
requested executed ✓ no data leaves your network

Identity · usage · billing

Every AI decision belongs to a customer.

Control, measure, and bill AI by who actually used it.

LIVE LEDGER

CUSTOMER

cust_acme

Acme Inc.

USAGE

1,284

requests

SPEND

$42.18

this month

BILLING

Stripe

Synced

Last updated 2 minutes ago

Impact across your product

See the effect across every request.

Vibe records what your app requested and what actually ran — so you can measure the real impact.

One call. Identity in code. Console shapes execution.

Pass identity (your customer) and workflow. UsageFlow scopes the ledger, merges Console policy at runtime, and meters what actually executes — no redeploy.

FAQ

Straight answers on Vibe SDK, gateways, and free tier.

Vibe SDK is a minimal AI-facing API (usage.chat(), usage.prompt()) that routes calls through UsageFlow's execution plan. You pass identity, workflow, and optional generation settings in code; policies in Console control models, tools, budgets, and how execution runs.
Gateways sit on the provider HTTP path. Vibe SDK runs inside your app, merges Console policy with your intent, transforms execution when needed, meters usage, and can sync executed usage to Stripe — on workflow and identity dimensions, not just API keys.
No. Policy edits in UsageFlow Console apply on the next call. Traces show requested vs executed settings and the reason for any transform.
Free includes one workflow with live policy (basic model ladder and Console edits without redeploy). Paid tiers add more workflows, tool gates, generation caps, and Stripe sync.
V0 targets OpenAI and Anthropic through a unified API. For full provider-specific surfaces, use usage.wrap(nativeClient) alongside the unified path.
Yes. Pass your customer or tenant id as identity in usage.chat(). UsageFlow scopes a ledger to that identity and can report (Console / Stripe) on workflow and identity dimensions — so spend reads in language your customers understand, not just API keys.
Built for teams shipping AI in production.

Write your AI app once.

Change models, caps, and tools from UsageFlow Console — no redeploy. Free includes one workflow with live policy.

Read Documentation

TypeScript, Python & Go • OpenAI + Anthropic •