One key to launch reliable AI features.
Token101 brings model access, routing resilience, metered billing, and usage visibility into one gateway, so your team can ship AI features without managing every provider separately.
Choose the integration path that fits your product
Start with the interface your product already uses, then add chat, reasoning, image generation, token planning, and model discovery from the same Token101 account.
Claude-style chat
Run Claude Code, long-context conversations, and streaming text through Token101 with minimal integration change.
/api/v1/messagesOpenAI-style chat
Keep existing OpenAI-compatible SDKs and agent frameworks while moving billing and model access into Token101.
/api/v1/chat/completionsReasoning workflows
Use a Responses-style route for multi-step reasoning, tool use, and newer application flows.
/api/v1/responsesImage generation
Offer image generation from the same account, with unified usage visibility and billing records.
/api/v1/images/generationsCost planning
Estimate tokens before a request so teams can control budget, trim context, and avoid surprise spend.
/api/v1/messages/count_tokensFrom account readiness to first business result
Token101 keeps the operational steps visible: create a key, confirm balance, choose the right model route, and get a successful response without hidden setup work.
Get a production key
Create a server-side key after signing in, then keep it in your backend or deployment secrets.
Create API keyMake spend predictable
Activate PAYG or a package before paid calls, and keep failed unsettled requests out of final spend.
Choose pricingShip the first feature
Pick the route your app already understands, copy a sample, and swap in your key and model ID.
Start quickstartSecurity note: never ship API keys to the browser
Examples use placeholders only. Real keys should be loaded from server environment variables and kept out of frontend bundles, logs, screenshots, and repositories.
TOKEN101_API_KEY=sk-...Send requests directly to the full endpoint URL — no SDK required.
curl https://token.ppthub.shop/api/v1/messages \
-X POST \
-H "Content-Type: application/json" \
-H "Authorization: Bearer sk-..." \
-H "Anthropic-Version: 2023-06-01" \
-d '{
"model": "claude-opus-4-8",
"max_tokens": 256,
"messages": [
{
"role": "user",
"content": "Summarize the benefits of a unified AI API in one paragraph."
}
]
}'Put leading AI models behind one commercial account
One Token101 account currently exposes 21 public model options across major providers, with one billing and usage trail.
Recommended launch models
Examples use the current model IDs we recommend for launch validation, keeping sales pages and integration docs aligned.
claude-opus-4-8gpt-5.4Know price and capability before you call
Model price, context window, supported modalities, and runtime status can be checked before a customer workload goes live.
Every paid call is protected before it reaches a provider
Token101 is more than a proxy. Each request passes access, policy, billing, usage, and fallback checks so teams can operate AI features with clearer control.
Access control
Missing, revoked, or insufficient keys return explicit errors before traffic reaches any upstream provider.
Model eligibility
Runtime disablement, account allowlists, and model availability are checked before routing.
Spend protection
Inactive PAYG, insufficient balance, or settlement failures return actionable reasons instead of silent cost drift.
Usage accountability
Successful requests link usage and credit transactions by request ID; failed requests are not finalized as spend.
Fallback routing
When providers rate-limit, timeout, or fail, Token101 can try backup channels and preserve debugging signals.
Turn evaluation questions into implementation answers
Authentication, costs, errors, limits, and tool integrations are organized around the questions teams ask before they trust an API in production.
Quickstart
Create a key, configure env vars, and send the first request.
API Reference
Compare available integration paths, request formats, and production routes.
Authentication
Understand Bearer tokens, key storage, and auth errors.
Responses API
Build reasoning workflows with the newer OpenAI-style interface.
Images API
Call image models and understand image billing semantics.
Error format
Handle 401, 402, 403, 429, and upstream failures.
Rate limits
Understand account protection, rolling windows, and response headers.
Billing
Learn PAYG, packages, discounts, and failed-request settlement.
Streaming
Build SSE streaming UX and handle client cancellation.
Tool calling
Prepare structured requests for agents and functions.
Codex integration
Use Token101 in Codex CLI and developer workflows.
Claude Code integration
Connect Claude Code through the Anthropic-compatible entry point.