v1.2.0Major Release
August 2026Introducing Vajram KN-1 Bilingual Inference & Neural Streaming
We are thrilled to launch the preview for Vajram KN-1, our high-concurrency bilingual Kannada & English foundation model adapter.
Sub-200ms time-to-first-token streaming architecture. Zero-shot and few-shot Kannada idiom & vachana preservation. Integrated client-side session caching with zero server logging. Enhanced markdown formatting with code block copy support. Autonomous Workflow Orchestration & Tool Calling
Added automated multi-step reasoning capabilities allowing agents to orchestrate scheduling, CRM syncs, and database lookups.
Structured function calling for calendar and communication APIs. Optimized prompt routing between small and large model tiers. Full support for custom enterprise rate limits and usage metering. LangBrains AI Platform Genesis
Official release of the LangBrains AI platform foundation and development sandbox.
Provider-neutral LLM runtime interface. Authentic clean light and dark theme design system. Comprehensive bilingual evaluation benchmarks.