AI Gateway
One connection
AI API Infrastructure
Access GPT, Gemini and Kimi through Zaitro API. Manage credits, API keys and usage from one clean developer platform.
THE ZAITRO CAPABILITY COLLECTION — 06
ONE PLATFORM. CONNECTED POSSIBILITIES.
From the first API key to a clear view of usage. Discover the pieces that bring your AI products together.
One connection
Connected possibilities
Access, in your hands
One balance
See your usage
Built for developers
One connection
Bring requests and model selection into one developer workflow.
FROM REQUEST TO RESPONSE
Your application connects through Zaitro API. Explore the request flow across GPT, Gemini and Kimi.
Product concepts and illustrative interfaces.POST /chat/completionsmodel: gpt-5.4gpt-5.4gemini-2.5-flashkimi-k2.7-codeBuilt for developers shipping AI products.
Model access
Three provider ecosystems behind a single API surface, a single key and a single balance.
Model explorer
38 illustrative model entries with example capabilities, context windows and API ids.
gpt-5.1gemini-3.8-flashkimi-k2.7-code-highspeedIllustrative catalog. Account models and prices are listed in your Console; upstream availability is not verified by this preview.
| Model | Provider | Capabilities | Context | Status | API ID |
|---|---|---|---|---|---|
| OpenAI | ReasoningCodingLong context | 1M | Available | gpt-5.4 | |
| Gemini | FastVisionLong context | 1M | Available | gemini-2.5-flash | |
| Kimi | ReasoningCodingLong context | 1M | Available | kimi-k3 | |
| OpenAI | FastLong context | 1M | Available | gpt-5.6-luna | |
| Gemini | FastLong context | 1M | Available | gemini-2.5-flash-lite | |
| Kimi | ReasoningCoding | 256K | Available | kimi-k2.6 | |
| OpenAI | ReasoningLong context | 1M | Available | gpt-5.6-sol | |
| Gemini | ReasoningVisionLong context | 1M | Available | gemini-2.5-pro | |
| Kimi | Coding | 256K | Available | kimi-k2.7-code |
Credits
One balance, drawn down by usage. The console shows what you have, what you have spent and what is left.
Add credit to a single Zaitro API balance. No subscription, no monthly minimum.
Spend that balance across GPT, Gemini and Kimi. Switching model does not mean switching account.
Requests, tokens and credits for every model and every key, in one view.
Wallet balance
Illustrative
$808.32
Available credits
Purchased
$1,000.00
Bonus
$350.00
Used today
$18.20
Estimated remaining
~40 days
At the last 7 days' rate
Last 7 days
73,736 requests
Top-up bonus
Eligible top-ups receive 35% extra credits on top of the amount you add. Move the slider to see what a top-up is worth.
Eligible top-ups of $20 or more receive 35% extra credits. Bonus credits are tracked separately from purchased credits in the console.
You receive
Bonus availability and promotional rates may vary. Bonus credits apply to eligible top-ups.
Why Zaitro API
The platform handles accounts, keys and accounting so your application only has to make a request.
01
Manage supported AI usage through a single Zaitro API balance rather than a separate account per provider.
02
Create, rotate and revoke keys from one console, with a spending limit on each.
03
Token consumption, request counts and credit spend, broken down by model and by key.
04
Reach supported GPT, Gemini and Kimi models without managing separate balances inside the Zaitro API platform.
Developers
Zaitro API speaks a chat-completions schema. Point your client at the base URL, pass a model id, and carry on.
curl https://api.zaitroapi.com/v1/chat/completions \ -H "Authorization: Bearer $ZAITRO_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-5.4", "messages": [ { "role": "user", "content": "Explain this report." } ] }'The API host is connected. Model generation depends on account access.
{ "id": "chatcmpl_7Qa4Kp", "model": "gpt-5.4", "choices": [ { "index": 0, "message": { "role": "assistant", "content": "The report covers three quarters..." }, "finish_reason": "stop" } ], "usage": { "prompt_tokens": 842, "completion_tokens": 311, "total_tokens": 1153 }}What that request costs
Deducted from your balance and attributed to the key that made the call.
Model switching
Your application talks to one base URL with one key. Choosing a different model means changing a string.
Unchangedhttps://api.zaitroapi.com/v1const res = await client.chat.completions.create({ model: "gpt-5.1", messages,});Same client, same request shape, same key. Only the model id moves. The exact request schema is confirmed against the backend before launch.
Playground
Pick a provider, pick a model, send a prompt. The console playground runs against your own balance.
Conversation
gpt-5.1You
Three themes account for most of the quarter: 1. Billing confusion after the plan change (41% of tickets) 2. Slow exports on reports over 50k rows (23%) 3. SSO setup for new workspaces (16%) The remaining 20% are one-off questions with no clear pattern.
Run the sample to see its usage.
Usage analytics
Requests, tokens and spend, broken down by provider, by model and by key.
API requests
73.7K
Across 7 points
Tokens
58.57M
35.70M in · 22.87M out
Credits used
$139.92
Deducted from balance
Error rate
0.49%
364 of 73,736
Volume by interval
Figures shown are sample account data for interface design. Live values come from your own usage.
Share of credits spent
Input plus output per interval
Credits spent, last 30 days
Figures shown are sample account data for interface design. Live values come from your own usage. See the console at /console.
API keys
Separate keys per environment, a spending limit on each, and one place to rotate or revoke.
API keys · Illustrative
| Key name | Created | Last used | Usage · 30d | Monthly limit | Status | Actions |
|---|---|---|---|---|---|---|
Production zai_live_7Qa4, remainder hidden | 14 Jun 2026 | 2 Oct 2026 | 24,180 | $400.00 | Active | Manage |
Development zai_test_3Kp9, remainder hidden | 14 Jun 2026 | 1 Oct 2026 | 3,942 | $50.00 | Active | Manage |
Internal Tool zai_live_1Rb6, remainder hidden | 2 Aug 2026 | 27 Sept 2026 | 1,548 | No limit | Active | Manage |
Prototype (rotated) zai_live_9Zd2, remainder hidden | 3 May 2026 | 1 Aug 2026 | 0 | No limit | Revoked | Archived |
For developers
Anything that makes a chat-completions request, from a single feature to a fleet of agents.
Product features built on chat completions.
Review, refactor and generation inside a developer workflow.
Summaries, extraction and classification over company data.
Scheduled jobs that read, decide and write back.
Multi-step tool use where model choice changes per step.
Batch evaluation across several models at once.
Triage, drafting and deflection on inbound tickets.
Enterprise
One account, many developers, and a clear view of what each of them is spending.
One balance for the whole team, topped up once.
A key per environment or per service, revocable in one click.
A monthly spending cap on any key.
Spend and volume by model, by key and by provider.
Separate projects with their own keys and budgets.
Who created, rotated or revoked which key, and when.
Consolidated invoicing and purchase orders.
Named contact and response commitments.
How it works
01
Email or Google. No sales call, no waiting list.
02
Prepaid balance, a key per environment.
03
Pass a model id and make your first request.
Documentation
Every endpoint documented with a copyable example and the parameters spelled out inline.
Create a key in the console, export it, and make your first request. Any OpenAI-compatible client works by changing the base URL.
export ZAITRO_API_KEY="zai_live_..." curl https://api.zaitroapi.com/v1/chat/completions \ -H "Authorization: Bearer $ZAITRO_API_KEY" \ -H "Content-Type: application/json" \ -d '{ "model": "gpt-5.4", "messages": [{"role":"user","content":"Hello"}] }'modelstringRequiredmessagesarrayRequiredtemperaturenumbermax_tokensintegerFAQ
Bracketed answers are policy decisions for the business to confirm before launch.
Start with Zaitro API and manage your supported AI usage from one developer platform.