Tokensor¶
Tokensor is a developer platform that sits in front of your LLM providers. It enforces per-account token quotas, gives you transparent time-series cost analytics, and lets you connect agent frameworks so you can track costs and detect inefficiencies in real time.
What you get¶
- Transparent proxies — route OpenAI, Anthropic, and OpenCode calls through Tokensor to record every call and its cost.
- Quota enforcement — per-account token quotas and daily spend ceilings, enforced before the request reaches your provider.
- Cost analytics — daily totals, cost-by-model, spend summary, and per-agent cost rollups.
- Agent cost intelligence — ingest OpenTelemetry traces from agent frameworks (OpenLLMetry/Traceloop, LangChain, agentgateway) to detect runaway loops, redundant tool calls, and context bloat.
- Billing tiers — quota-based free tier with paid upgrades.
Quick links¶
- Setup — create your account, get your API key, and connect a provider.
- API Reference — endpoints, authentication, and examples.
- Integrations — OpenAI, Anthropic, OpenCode, and agent frameworks.
How it works¶
All traffic passes through Tokensor's proxies so every call is recorded, quota-checked, and attributed to the correct account.
Getting started¶
- Sign in at tokensor.dev and copy your Tokensor API key.
- Connect a provider key in onboarding or Settings.
- Point your app at Tokensor instead of the provider directly.
- Watch costs appear in the dashboard within ~60 seconds.