Resources

Changelog

Changes to the Laras Router API, models, pricing, and documentation.

On this page

29 October 2026 — Admin fee is 1%

The admin fee on top-ups rises from 0.70% to 1% of the amount.

  • Wallet. The fee is still itemised on every invoice before you pay. Invoices issued before this change keep the fee they were quoted at.

28 October 2026 — Anthropic Messages API and a $1 minimum top-up

Claude Code and the Anthropic SDKs can connect, and you can start with a smaller top-up.

  • API. POST /v1/messages, the Anthropic Messages API, at https://api.larasrouter.com. Streaming, tool use, and thinking blocks are supported. It runs the same model at the same prices and limits as chat completions and is billed from the usage the model reports.
  • Wallet. The minimum top-up is now $1, down from $3. The maximum stays $100.
  • Docs. A Claude Code guide and a Messages API reference.

Coming next: the OpenAI Responses API (Codex CLI).

12 October 2026 — Launch

Laras Router is open to everyone.

  • API. OpenAI-compatible POST /v1/chat/completions and GET /v1/models at https://api.larasrouter.com/v1, with streaming, tool calling, JSON output, and reasoning content.
  • Model. deepseek-flash (DeepSeek-V4.1-Flash): 1M-token context, up to 384K output tokens. $0.010 / $0.400 / $1.600 per 1M tokens for cache-hit input, cache-miss input, and output.
  • Wallet. Prepaid USD balance, topped up in Rupiah with QRIS from $3 to $100, with a 0.70% admin fee shown on every invoice.
  • API keys. Per-key limits of 300 requests per minute and 16 concurrent requests, and optional monthly budgets.
  • Docs. This site, in English and Bahasa Indonesia.

Coming next: support for the Anthropic Messages format (Claude Code) and the OpenAI Responses API (Codex CLI).