Light API Platform

Unified AI Chat, Terminal Assistant and Model API Gateway

Light API brings together web chat, a terminal assistant and an OpenAI/Anthropic-compatible API gateway backed by server-side accounting. Start with the free daily allowance in your browser or connect a capped personal API key to VS Code, Cursor, Claude Code and OpenCode.

Models and Pricing App Connection API Documentation FAQ Open Web Workspace

Connected models14 of 20Brands in catalog: 10

Free chat allowance10 / 50Requests/day (guest / signed-in), models: 12

BillingInput / output / cachePer-model rates per 1M tokens

Gateway protocolsOpenAI + AnthropicSSE streaming, key caps & quota reserve

Rate status: All active model rates in the catalog are marked as verified in the server configuration.

How the Platform Works

All rate limits, quota reservations and key controls are enforced on the server and inside the gateway, with transparent usage settlement.

Web Chat & Light AI Terminal

Use up to 10 free messages per day without an account, or sign in with Telegram or a password for up to 50 free requests per day (up to 2024 input tokens, 512 output tokens, and memory of the last 5 messages).

Read about free limits

Two Account Types

Signing in with Telegram opens the shared wallet linked to the Telegram bot. Registering with a username and password on the website creates an independent account with its own separate wallet balance.

Account setup guide

Capped & Expiring API Keys

Create personal API keys in your account, select allowed models, set a USD spending cap and expiration (7, 30, 90, 365 days or no expiration), and replace or revoke keys at any time.

Manage API keys

Desktop Installers & Standard Protocols

The gateway supports OpenAI (/v1/chat/completions) and Anthropic (/v1/messages) protocols, with ready setup archives for Windows and macOS.

Explore API docs

Connected Models in Catalog

Below are the models currently active in the server configuration. See the full catalog page for all variants, planned models and billing modes.

OpenAI

ConnectedFree chat

GPT 5.6 Sol

gpt-5.6-sol

GPT 5.6 Sol by OpenAI.

Input: $0.02
Output: $0.10
Cache read: $0.002
Cache write: $0.025

Per 1M tokens · Verified rate

OpenAI

ConnectedFree chat

GPT 6 Astra

gpt-6-astra

OpenAI model for complex work: detailed reasoning, analysis and coding.

Input: $0.05
Output: $0.25
Cache read: $0.005
Cache write: $0.0625

Per 1M tokens · Verified rate

Anthropic

ConnectedFree chat

Claude Sonnet 5.5

claude-sonnet-5-5

Claude Sonnet 5.5 by Anthropic.

Input: $0.05
Output: $0.25
Cache read: $0.0025
Cache write: $0.0625

Per 1M tokens · Verified rate

Anthropic

Connected

Opus 5.5 - supplier 1

claude-opus-5-5-provider-1

Opus 5.5 by Anthropic through supplier 1.

Input: $0.08
Output: $0.40
Cache read: $0.004
Cache write: $0.10

Per 1M tokens · Verified rate

Z.ai

ConnectedFree chat

GLM 5.3

glm-5.3

GLM 5.3 by Z.ai for conversation, text and code.

Input: $0.021
Output: $0.066
Cache read: $0.0039
Cache write: $0.021

Per 1M tokens · Verified rate

Moonshot AI

ConnectedFree chat

Kimi K3

kimi-k3

Kimi K3 by Moonshot AI. Another option for writing, analysis and programming.

Input: $0.036
Output: $0.18
Cache read: $0.0036
Cache write: $0.036

Per 1M tokens · Verified rate

DeepSeek

ConnectedFree chat

DeepSeek V4 Pro

deepseek-v4-pro

The Pro variant of DeepSeek for text, code and detailed tasks. Compare it on your own workload.

Input: $0.00264
Output: $0.00792
Cache read: $0.000088
Cache write: $0.00264

Per 1M tokens · Verified rate

DeepSeek

ConnectedFree chat

DeepSeek V4.1 Flash

deepseek-v4.1-flash

A fast model for everyday tasks: explaining a topic, drafting text, brainstorming ideas or helping with simple code. A convenient starting point for AI.

Input: $0.0054
Output: $0.0216
Cache read: $0.000108
Cache write: $0.0054

Per 1M tokens · Verified rate

Qwen

ConnectedFree chat

Qwen 3.8 Flash

qwen3.8-flash

Qwen 3.8 Flash by Qwen.

Input: $0.0064
Output: $0.0188
Cache read: $0.00064
Cache write: $0.0064

Per 1M tokens · Verified rate

MiMo

ConnectedFree chat

MiMo V2.6 Pro

mimo-v2.6-pro

MiMo V2.6 Pro by MiMo.

Input: $0.01305
Output: $0.0261
Cache read: $0.000108
Cache write: $0.01305

Per 1M tokens · Verified rate

xAI

ConnectedFree chat

Grok 4.7

grok-4.7

Grok 4.7 by xAI.

Input: $0.014
Output: $0.042
Cache read: $0.0035
Cache write: $0.014

Per 1M tokens · Verified rate

Google

ConnectedFree chat

Gemini 3.8 Flash

gemini-3.8-flash

Gemini 3.8 Flash by Google.

Input: $0.0225
Output: $0.1125
Cache read: $0.00225
Cache write: $0.0225

Per 1M tokens · Verified rate

ConnectedFree chat

HY4 Preview

hy4-preview

HY4 Preview by HY.

Input: $0.00834
Output: $0.02501
Cache read: $0.00042
Cache write: $0.00834

Per 1M tokens · Verified rate

Google

Connected

Gemini 4 Argon

gemini-4-argon

Gemini 4 Argon by Google.

Input: $0.20
Output: $1.00
Cache read: $0.01
Cache write: $0.01

Per 1M tokens · Verified rate

View Full Catalog & Rates

Quota Reservation & Referral Program

Pre-Request Quota Reservation

Before dispatching a paid text request, the gateway verifies both your key spending cap and wallet balance, reserves the upper-bound token cost and automatically refunds any unused reservation once the response finishes.

10% Referral Commission

You earn 10% from confirmed genuine top-ups made by users who join through your referral link. Test credits do not generate commission, refunds adjust rewards, and payouts are processed manually through support.