Platform Changelog
A chronological record of updates to the Mira platform, API, and developer tools. New entries are added at the top.
Subscribe to updates at platform.vmira.ai to receive notifications about new releases and API changes.
April 2026
Documentation & Mira Code v0.1.0
- Launched documentation site platform.vmira.ai/docs with full bilingual support (EN/RU)
- Released Mira Code v0.1.0 — stable CLI version for developers
- Added SDKs, security, glossary, and changelog documentation pages
- Improved documentation navigation with sections and search
- Added integration examples for LangChain, LlamaIndex, and major SDKs
March 2026
mira-max & Batch Processing
- Launched mira-max model — most capable model with 128K token context window
- Added Batch Processing API for bulk requests at 50% discount
- Increased mira-pro context window to 64K tokens
- New /v1/batches endpoint for creating and managing batch jobs
- Improved streaming performance — reduced time-to-first-token by 35%
- Added structured outputs (JSON mode) support for all models
February 2026
Device Code Auth & Mira Code Beta
- Implemented device code flow authentication for CLI tools
- Launched Mira Code Beta — AI-powered development assistant in the terminal
- Improved API rate limits: increased RPM for Pro and Max plans
- Added support for scoped API keys with limited permissions
- New API usage dashboard with detailed analytics
- Fixed rare timeout errors during long streaming responses
January 2026
mira-pro, Vision & Tool Use
- Launched mira-pro model — advanced model for professional tasks
- Added Vision (image analysis) support for mira-pro and mira-max
- Implemented Tool Use / Function Calling for automatic integration with external APIs
- Multi-turn conversation support with message history up to 64K tokens
- Added /v1/embeddings endpoint for text vector embedding generation
- Introduced Pro and Max subscription plans with extended limits
December 2025
Platform Launch
- Launched the Mira platform — api.vmira.ai and platform.vmira.ai
- Available models: mira (general-purpose), mira-pro (with thinking mode), and mira-max (maximum capabilities)
- OpenAI-compatible API: /v1/chat/completions, /v1/models
- Full streaming support via Server-Sent Events (SSE)
- Registration and API key management through the dashboard
- API documentation and code examples for Python, JavaScript, and cURL
- Free tier: 20 requests per day to explore the platform
Upcoming Updates
We are continuously working on improving the platform. Here is what is planned for the near future:
- Fine-tuning API — Ability to fine-tune models on your data through the API.
- Webhooks — Notifications for batch job completion and other events.
- Assistants API — Create persistent AI assistants with file storage and tools.
- Model expansion — New specialized models for coding, analysis, and generation.