InsForge — backend for AI coding agents

★ 13K

InsForge is an open-source backend platform for agentic coding that gives an AI agent a database, authentication, file storage, edge functions, a model gateway, compute and site deployment through a single interface. Reach for it when you want an agent to ship a full-stack app end to end without you wiring up Postgres, S3-style storage, auth and hosting separately; or when the agent needs to deploy edge functions, run database migrations, inspect bucket contents and read runtime logs on its own. Inside are Postgres, S3-compatible storage (files sit on the local filesystem by default, with MinIO or RustFS as an optional overlay), user and session management, an OpenAI-compatible gateway across several LLM providers, serverless edge functions, site build and deployment, and long-running container services in compute (private preview). Agents connect through an MCP server, available both self-hosted and in the cloud, or through a CLI paired with Skills, cloud only. Self-hosting runs on Docker Compose (Docker with Compose v2 is required) through the setup.sh script, and the dashboard opens at localhost:7130; the cloud version lives at insforge.dev. Licensed under Apache-2.0. What sets it apart is that the agent gets a live, operable backend with docs, schemas and logs rather than just generating code. A good fit for teams building apps with AI agents; the CLI mode needs the cloud version.