InsForge is an all-in-one, open-source backend platform for agentic coding that gives a coding agent database, auth, storage, compute, hosting, and an AI gateway so it can ship full-stack apps end to end.

What it is

InsForge is a TypeScript backend platform, licensed Apache-2.0, that packages authentication, a Postgres relational database, S3-compatible file storage, a model gateway, edge functions, compute, and site deployment into one system. Coding agents reach it through one of two interfaces: an MCP Server, available self-hosted and cloud, which exposes InsForge's operations as tools any MCP-compatible agent can call, or a CLI paired with Skills, cloud only, which agents invoke directly from the terminal. Through those interfaces an agent reads backend context and state — documentation, schemas, metadata such as deployed functions, bucket contents and auth config, and runtime logs — and configures primitives directly, deploying edge functions, running database migrations, creating storage buckets, and setting up auth providers.

The problem it solves is the assembly work that usually sits between an agent and a working backend. Each primitive would otherwise be a separate service wired together by hand and described to the agent through documentation the agent cannot verify against reality, leaving it unable to confirm what it actually built or to debug what broke. InsForge replaces that hand-assembled collection of backend services with a single platform the agent can operate and inspect, so it can write code, verify the result, and diagnose failures against live runtime logs rather than assumptions.

Key capabilities

• MCP Server interface, self-hosted and cloud, exposing InsForge operations as tools for any MCP-compatible coding agent.
• CLI + Skills interface, cloud only, letting agents drive the backend from the terminal.
• Postgres relational database, with pgvector and vector and embedding support for AI workloads.
• S3-compatible file storage, including storage buckets the agent can create and inspect.
• OpenAI-compatible Model Gateway offering one API across multiple LLM providers.
• Edge Functions for serverless code on the edge, plus site build and deployment.
• Compute, in private preview, for long-running container services, alongside OAuth2 authentication and realtime features.

Who uses it and how

• Developers running agentic coding tools that speak MCP, where the agent calls InsForge tools to run migrations, deploy functions, and read schemas without leaving its loop.
• Teams self-hosting the full stack: the setup script writes ~/insforge/.env with generated secrets, then docker compose up -d starts the platform from that directory.
• Cloud users at insforge.dev who want managed backend operations, with CLI + Skills for terminal-driven agents.
• AI application builders using the Postgres database with pgvector for embeddings and vectors, and the model gateway instead of integrating each LLM provider separately.
• Teams evaluating long-running container workloads through Compute while it remains in private preview.

Getting started

Cloud-hosted, the platform runs at insforge.dev. Self-hosted, the README's path is Docker with Compose v2: run the deploy/setup.sh script to generate ~/insforge/.env and fetch the stack's files, edit API_BASE_URL and VITE_API_BASE_URL, then run docker compose up -d.

How it compares

The facts provide no list of paid products this project replaces, and none of the represented tools are named as close equivalents. On the evidence available, InsForge stands alone in this registry: an open-source backend platform whose distinguishing feature is that its primary operator is a coding agent rather than a human clicking through a dashboard.

When to use it — and when not to

A self-hoster must operate the whole Docker Compose stack and own its secrets: JWT_SECRET, ENCRYPTION_KEY, POSTGRES_PASSWORD, and ROOT_ADMIN_PASSWORD, plus ACCESS_API_KEY and ACCESS_ANON_KEY if the keys should be known rather than backend-generated. That is real operational work — Postgres, storage, and the surrounding services are yours to run and secure. The source-build path (docker-compose.prod.yml) generates nothing and falls back to published defaults for any unset variable, so exposing it without setting secrets first is unsafe. Anyone not using an agentic coding workflow, or unwilling to run Docker, should look elsewhere; Compute is still private preview, and the README is quickstart-oriented, with lighter reference material than a mature backend platform would carry.

graph TB

    subgraph TOP[" "]
        AG[AI Coding Agents]
    end

    subgraph MID[" "]
        SL[InsForge]
    end

    AG --> SL

    SL --> AUTH[Authentication]
    SL --> DB[Database]
    SL --> ST[Storage]
    SL --> EF[Edge Functions]
    SL --> MG[Model Gateway]
    SL --> CP[Compute]
    SL --> DEP[Deployment]

    classDef bar fill:#0b0f14,stroke:#30363d,stroke-width:1px,color:#ffffff
    classDef card fill:#161b22,stroke:#30363d,stroke-width:1px,color:#ffffff

    class AG,SL bar
    class AUTH,DB,ST,EF,MG,CP,DEP card

    style TOP fill:transparent,stroke:transparent
    style MID fill:transparent,stroke:transparent

    linkStyle default stroke:#30363d,stroke-width:1px

curl -fsSL https://raw.githubusercontent.com/InsForge/InsForge/main/deploy/setup.sh | sh -s ~/insforge

cd ~/insforge
$EDITOR .env          # API_BASE_URL, VITE_API_BASE_URL — the URL browsers will use
docker compose up -d

git clone https://github.com/InsForge/InsForge.git
cd InsForge
cp .env.example .env
$EDITOR .env
docker compose -f docker-compose.prod.yml up

I'm using InsForge as my backend platform, call InsForge MCP's fetch-docs tool to learn about InsForge instructions.

curl -fsSL https://raw.githubusercontent.com/InsForge/InsForge/main/deploy/setup.sh | sh -s ~/project1
curl -fsSL https://raw.githubusercontent.com/InsForge/InsForge/main/deploy/setup.sh | sh -s ~/project2