Product Update: What Changed in January 2026

Highlights

We shipped clearer documentation entry points, better request cost visibility, and expanded fallback controls for production workloads.

Upcoming work focuses on status transparency, changelog hygiene, and guided integration templates.

What shipped

Improved documentation structure — The docs at aicredits.in/docs now have a clearer information hierarchy. Quickstart, API reference, authentication, and pricing each have dedicated pages with working code examples.

Per-request cost visibility — The usage dashboard now shows the INR cost for every individual request, broken down into input token cost, output token cost, and forex conversion. No more estimating from totals.

Expanded fallback controls — You can now configure fallback behaviour per API key. Set whether a key should fall back automatically on provider failure or surface the error to your application for custom handling.

Model routing safeguards — Added validation that catches common model ID mistakes (e.g., using claude-3-opus instead of claude-3-opus-20240229) and returns a helpful error message rather than a generic 404.

What's coming in February

• Semantic caching dashboard showing hit rate and estimated savings
• Webhook support for low-balance alerts
• Bulk API key management for team accounts
• Expanded model catalogue with DeepSeek R2 and Mistral Large 2

Follow the changelog at aicredits.in/changelog or subscribe to the newsletter for updates as they ship.