AIGet by Seneco
One interface for every model.
AIGet puts many AI models behind one OpenAI-compatible endpoint. It handles routing, provider failover, usage tracking and token-based billing from a single prepaid balance.
The gateway
One integration, the whole model ecosystem.
Stop wiring a separate SDK, key and invoice for every provider. AIGet accepts one OpenAI-compatible request and decides where it should go.
Chat, tools, images, PDFs and audio
One API surface for the capabilities models actually support. Unsupported models are filtered out before a request is sent.
Major providers, one catalog
Google, OpenAI, Anthropic, Meta, Mistral, DeepSeek, xAI, Qwen and more, listed with their published per-token prices.
Failover before the first token
If a provider rate-limits or fails, AIGet retries another candidate before anything reaches your application.
Prepaid, metered and traceable
A USD balance is charged per token actually used. Each charge records the request behind it.
curl https://api.aiget.tech/v1/chat/completions \ -H "Authorization: Bearer $AIGET_KEY" \ -H "Content-Type: application/json" \ -H "X-AIGet-Policy: balanced" \ -d '{ "model": "aiget/auto", "messages": [{ "role": "user", "content": "Hello" }], "stream": true }'Routing & policies
Pin a model, or let the router decide.
Send a specific model id when you need exact behaviour. Send aiget/auto when you want the gateway to choose by cost, quality, latency or a balanced mix.
- X-AIGet-Policy: cost
- X-AIGet-Policy: quality
- X-AIGet-Policy: latency
- X-AIGet-Policy: balanced
- Response: X-AIGet-Model
- Response: X-AIGet-Provider
import OpenAI from "openai"; const client = new OpenAI({ baseURL: "https://api.aiget.tech/v1", apiKey: process.env.AIGET_KEY,}); const completion = await client.chat.completions.create({ model: "deepseek-ai/deepseek-v4.1-flash-1017", messages: [{ role: "user", content: "Hello" }],});NoteModel ids, capabilities and prices change as providers update their catalogs. Use GET /v1/models or the AIGet model pages for the current list.
Billing
Prepaid, per token, no subscription.
Top up once and spend it whenever. The balance is held in USD, charged per token actually used, and credits do not expire.
- $5 minimum top-up
- Top-ups paid in IDR via Midtrans
- Exchange rate shown before payment
- No monthly minimum or platform fee
New accounts
- $5 free credit after email verification
- 2FA enabled and first API key created
- No credit card required
- Free models keep working at zero balance
AIGet's pricing is shown per model, per million input, output and cache-read tokens. The exact price is on each model page.
Keys & data
Built to keep your data yours.
AIGet documents its controls publicly: hashed keys, encrypted provider credentials, organisation-scoped requests and a clear prompt policy.
Keys are not readable
API keys are stored in a form that cannot be reversed and shown only once.
Provider credentials encrypted
Decryption material is kept separately from the stored credentials.
Prompts not stored for training
AIGet states prompts are not stored or used for training.
Organisation-scoped
A key only reaches its project; usage and balance stay isolated.
Traceable charges
Every balance change has a reason and a reference.
Restricted admin access
Admin tools require a secret token and a permitted origin.
AIGet questions
Practical answers.
For error codes, rate limits, streaming details and the full endpoint reference, use the AIGet documentation.
Is AIGet a model provider?
No. AIGet is a gateway in front of providers. It routes requests, handles failover, tracks usage and bills from one prepaid balance. The models themselves are served by the provider that answers.
Do I have to change my application?
Usually only the base URL and API key. Requests and responses follow OpenAI Chat Completions, so the OpenAI SDK, LangChain, the Vercel AI SDK and other compatible clients work without meaningful changes.
How do I choose a model?
Send a model id from GET /v1/models to pin it, or send aiget/auto and set X-AIGet-Policy to cost, quality, latency or balanced. The response headers report which model and provider actually answered.
What happens when a provider fails?
Rate limits and 5xx responses from a provider are retried automatically before anything is sent back to you. If every candidate fails, the API returns all_providers_failed.
Are prompts stored or used for training?
AIGet's published policy is that prompts are not stored or used for training. You choose which providers may receive your data, and usage and charges remain traceable in the dashboard.
AIGet · live
One endpoint. Every model.
Create an account, verify your email, enable 2FA and generate a key. The $5 credit is added automatically.
No credit card required. AIGet is live; SentinelX Cloud is still in development.