Changelog
New features, improvements, and fixes shipped to InsForge.
Site Deployment (Vercel)
InsForge now supports site deployment via the Vercel Platform API, enabling fully automated build and deployment workflows.
Expanded AI Model Support in Model Gateway
InsForge Model Gateway now supports many additional open source and SOTA AI models, including Minimax, DeepSeek, Grok, GLMs, and more.
PDF Ingestion and Web Search in Model Gateway
InsForge Model Gateway now supports PDF ingestion and web search, enabling agents to read PDFs and search the web within a single workflow.
RPC Function Support
You can now call database RPC functions directly using the InsForge SDK, allowing applications to run database logic without custom API routes.
Audio Input Support in Model Gateway
InsForge Model Gateway now supports audio input modalities for LLMs, enabling translator apps and voice-driven agent workflows.
Kiro and Antigravity MCP Installers
InsForge now provides official MCP installers for Kiro and Antigravity, allowing Kiro and Antigravity developers to connect to InsForge by running a single command.
New Landing Page
And a super cool new landing page!
Realtime Module
We launched a new Realtime module built on WebSockets with a pub/sub architecture. It supports channels for live updates, message history for state syncing, and RLS-based permissions out of the box—making it well suited for real-time apps, games, chat, and collaborative experiences.
View Realtime SDK documentation
Apple OAuth
Apple OAuth is now available in InsForge Auth, as requested by the community. You can enable Sign in with Apple in a few clicks and ship apps with Apple login without manual OAuth setup.
Alipay Payments
Alipay is now supported for InsForge subscriptions, so you can subscribe without a credit card. This removes a major payment barrier for users in the APAC region.
More Models in the Model Gateway
We have added more popular and state-of-the-art models to the InsForge Model Gateway so you can choose the right model for each use case:
- GPT 5.2 - Stronger reasoning and longer context
- DeepSeek V3 - High performance at a lower cost
- Gemini 3 Flash - Low latency real-time inference
More models are coming soon.
Multi-Region Deployments
Deploy your project to the appropriate region for improved latency and compliance. We currently support us-east, us-west, eu-central and ap-southeast regions. More regions will be added based on demand.
Postgres Connection
Connect to your project's Postgres database and query it directly from your database clients. This also allows you to integrate InsForge with external services and tools like n8n.
