One API key.
Every frontier model.
A unified OpenAI-compatible gateway to GPT, Claude, Gemini, and DeepSeek. Swap one base URL, keep your existing SDK code — no rewrites, no vendor lock-in.
No credit card required · 1M free tokens on signup
from openai import OpenAI
client = OpenAI(
api_key="sk-tokenso-xxxxxxxxxxxxxxxx",
base_url="https://api.tokenso.top/v1", # only this line changes
)
resp = client.chat.completions.create(
model="gpt-4o",
messages=[{"role": "user", "content": "Explain zero-copy IO in one sentence."}],
)
print(resp.choices[0].message.content)Built for teams shipping production AI
Drop-in compatibility, intelligent routing, and transparent per-token billing — without touching your application code.
OpenAI-compatible
Point your existing OpenAI SDK at our endpoint and it just works. Same request shape, same response schema, same streaming behaviour.
Smart routing
Requests are served from the nearest healthy upstream. Automatic failover between providers keeps your service online during outages.
Transparent pricing
Pay strictly for tokens consumed, with no subscription floor. Every request is itemised in your usage dashboard.
High concurrency
Burst to thousands of concurrent streams. No artificial rate limits on paid tiers, with per-key quotas you control.
Key management
Issue scoped keys per project or per teammate, set spend caps, and revoke instantly. Full audit log on every call.
Streaming first
Server-sent events are passed through untouched. Time-to-first-token stays low even under sustained load.
Frontier models, one endpoint
All models are addressed with the standard model parameter. Pricing is per 1M tokens, blended input/output.
| Model ID | Provider | Context | Price / 1M | Status |
|---|---|---|---|---|
| gpt-4o | OpenAI | 128K | $2.50 | Popular |
| gpt-4o-mini | OpenAI | 128K | $0.15 | Stable |
| claude-sonnet-4 | Anthropic | 200K | $3.00 | New |
| claude-opus-4 | Anthropic | 200K | $15.00 | Stable |
| gemini-2.5-pro | 1M | $1.25 | New | |
| gemini-2.5-flash | 1M | $0.30 | Stable | |
| deepseek-chat | DeepSeek | 128K | $0.27 | Stable |
| qwen-max | Alibaba | 32K | $0.40 | Stable |
Pricing shown in USD. Volume discounts available above 500M tokens per month — contact sales.
Pay for what you use
Start free. Scale when you ship. Every plan includes full model access and the same low-latency infrastructure.
Free
For evaluation and side projects
- 1M tokens included
- All models unlocked
- 5 requests / minute
- Community support
Developer
For production apps and small teams
- Unlimited throughput
- Priority routing
- 10 API keys, spend caps
- Usage analytics dashboard
- Email support, 24h
Enterprise
For high-volume and regulated workloads
- Dedicated capacity
- Volume token pricing
- SSO & team roles
- Custom data retention
- Slack channel, 1h SLA
Live in under two minutes
Three steps. No migration, no dependency changes, no code rewrite.
Create an account
Sign up and generate an API key from the console. Keys are prefixed with sk- and can be scoped per project.
Change the base URL
Set base_url to our endpoint. Everything else — headers, payload, response schema — stays identical to OpenAI.
Send your first request
Run your existing code unchanged. Streaming, function calling, and vision inputs are all supported out of the box.
Questions, answered
In most cases, no. Any OpenAI-compatible SDK works by changing a single configuration value — the base URL.
GPT, Claude, Gemini, DeepSeek, Qwen, and Llama families are available through a single endpoint.
Billing is strictly per token consumed, split into input and output tokens at the rates listed above.
Health checks run continuously against every configured upstream. When a provider degrades, traffic is rerouted to the next healthy endpoint.
No. Request and response payloads are not used for model training.
Yes. Monthly invoices are generated automatically and are downloadable as PDF from the billing section of your console.
Start building in two minutes
Generate a key, change one line, and ship. Free tier included, no card required.