Allotly Documentation

Complete API documentation for Allotly — authentication, team management, budget controls, and AI model access for OpenAI, Anthropic, Gemini, and Azure.

What is Allotly

Allotly is the AI Spend Control Plane. It sits between your organization and AI providers like OpenAI, Anthropic, and Google Gemini, giving you complete visibility and control over who can access which models and how much they can spend.

Allotly Teams

Direct-access model. Members receive scoped API keys from their AI provider and call OpenAI, Anthropic, or Google through Allotly's unified proxy. Allotly tracks usage in real-time and enforces budgets per request.

Allotly Vouchers

Proxy-access model. Create voucher codes with budgets. Recipients get an Allotly API key and route requests through our OpenAI-compatible proxy, which enforces limits in real-time per request.

Quick Start (5-Minute Setup)

  1. Sign Up — Create your organization at /signup. You become the Root Admin automatically.
  2. Connect an AI Provider — Add your OpenAI, Anthropic, or Google API key. Keys are encrypted with AES-256-GCM.
  3. Create a Team — Go to Teams and create your first team.
  4. Add Members — Invite team members by email. Each member gets scoped API keys.
  5. Set Budgets — Assign per-member monthly budgets. Allotly will alert at 80% and 90%, and revoke keys at 100%.

Key Concepts

Root Admin
Full organizational control. Connects AI providers, creates Team Admins, manages billing and plan settings.
Team Admin
Manages one team. Adds members, sets per-member budgets, creates vouchers, views team analytics.
Member
End user. Receives API keys, tracks their own usage, operates within assigned budget limits.

Connecting Providers

Allotly supports OpenAI, Anthropic (Claude), Google Gemini, and Azure OpenAI. Navigate to AI Providers in your dashboard and enter your provider API key. The key is validated in real time and encrypted with AES-256-GCM before storage.

Proxy API Reference

The Allotly proxy is fully OpenAI-compatible. Replace your OpenAI base URL with your Allotly proxy URL and your API key with your Allotly member key. All OpenAI-SDK-compatible tools work without code changes.

Proxy endpoint: https://allotly.ai/v1

FAQ

What if Allotly goes down?

All requests route through Allotly's unified proxy, so they would be unavailable if the proxy is down. Service resumes automatically when the proxy recovers. We monitor uptime 24/7 and target 99.9% availability.

Do you store prompts?

Never. The proxy processes requests in-flight and only logs metadata: token counts, estimated costs, timestamps, model names, and status codes. Your prompts and AI responses are never written to disk.

How accurate are budgets?

Budget enforcement is real-time per-request. The proxy reserves budget before forwarding each request, then adjusts based on actual token usage from the provider response. Accuracy is within a fraction of a cent.

Can I use Allotly with LangChain, Cursor, or other tools?

Yes. The Allotly proxy is fully OpenAI-compatible, so any tool that supports custom OpenAI base URLs will work — including LangChain, Cursor, VS Code extensions, and any OpenAI SDK.

When should I use Teams vs Vouchers?

Use Teams for ongoing AI access to internal team members with monthly resetting budgets. Use Vouchers to distribute temporary, budget-capped AI access to people outside your organization.

What are the different voucher tiers?

Free plan: 1 active code, 2 redemptions, $5 max budget, 1-day expiry. Team plan: 5 codes per admin, 5 redemptions each, $20 max budget, 30-day expiry. Bundle ($10 one-time): 10 codes, 50 pooled redemptions, $25 max budget, no expiry — unused capacity carries over indefinitely.