GoFr is designed to simplify microservice development, with a key focus on Kubernetes deployment and out-of-the-box observability. While capable of building generic applications, microservices remain at its core.
- Simple API Syntax
- REST Standards by Default
- Configuration Management
- Observability (Logs, Traces, Metrics)
- Inbuilt Auth Middleware & Custom Middleware Support
- gRPC Support
- HTTP Service with Circuit Breaker Support
- Pub/Sub
- Health Check for All Datasources
- Database Migration
- Cron Jobs
- Support for Changing Log Level Without Restarting
- Swagger Rendering
- Abstracted File Systems
- Websockets
- LLM Integration (
app.AddLLM,ctx.LLM()) with Built-in Tracing, Token Metrics and Streaming - MCP Server (
app.EnableMCP) to Expose Read-Only Handlers as Tools for AI Agents
To get started with GoFr, add the following import to your code and use Go’s module support to automatically fetch dependencies:
import "gofr.dev/pkg/gofr"Alternatively, use the command:
go get -u gofr.dev/pkg/gofrHere's a simple example to get a GoFr application up and running:
package main
import "gofr.dev/pkg/gofr"
func main() {
app := gofr.New()
app.GET("/greet", func(ctx *gofr.Context) (any, error) {
return "Hello World!", nil
})
app.Run() // listens and serves on localhost:8000
}To run this code:
$ go run main.goVisit localhost:8000/greet to see the result.
GoFr treats LLMs and AI agents like any other datasource: one line to set them up, with tracing, metrics, logs and health checks wired in automatically.
app.AddLLM registers an OpenAI-compatible model (OpenAI, Groq, DeepSeek, Together, Ollama, or any endpoint via
BaseURL). Handlers reach it through ctx.LLM(), which also offers Chat, Stream and Embed.
package main
import (
"gofr.dev/pkg/gofr"
"gofr.dev/pkg/gofr/ai/llm"
)
func main() {
app := gofr.New()
// The API key is read from LLM_API_KEY, or GROQ_API_KEY for this provider.
app.AddLLM(&llm.Client{Provider: llm.Groq, Model: "llama-3.3-70b-versatile"})
app.POST("/summarize", func(ctx *gofr.Context) (any, error) {
var in struct {
Text string `json:"text"`
}
if err := ctx.Bind(&in); err != nil {
return nil, err
}
return ctx.LLM().Generate(ctx, "Summarize the following:\n"+in.Text)
})
app.Run()
}A model can also be configured without code, through LLM_PROVIDER, LLM_MODEL and LLM_API_KEY in configs/.env.
See the LLM guide.
app.EnableMCP serves your read-only handlers (GET, HEAD, OPTIONS and QUERY) as
Model Context Protocol tools, so MCP clients such as Claude Code or Claude Desktop
can discover and call them. Write handlers are never exposed. The MCP server listens on loopback at MCP_PORT
(default 8200; 0 turns the transport off).
app := gofr.New()
app.EnableMCP()
app.GET("/products/{id}", getProduct)
app.Run()ctx.LLM().Tools() gives a handler the same tools in-process, so you can build your own agent loop. See
Building AI Agents and the
using-ai example.
Explore a variety of ready-to-run examples in the GoFr examples directory.
- GoDoc: Official API documentation.
- GoFr Documentation: Comprehensive guides and resources.
- AI coding assistants: point your assistant at AGENTS.md for GoFr's conventions, and at llms.txt for an index of the docs.
Join Us in Making GoFr Better
Share your experience: If you’ve found GoFr helpful, consider writing a review or tutorial on platforms like Medium, Dev.to, or your personal blog. Your insights could help others get started faster!
Contribute to the project: Want to get involved? Check out our CONTRIBUTING.md guide to learn how you can contribute code, suggest improvements, or report issues.
To securely clone the GoFr repository, you can use HTTPS or SSH:
git clone https://github.com/gofr-dev/gofr.gitgit clone git@github.com:gofr-dev/gofr.gitIf your PR is merged, or if you contribute by writing articles or promoting GoFr, we invite you to fill out this form to claim your GoFr merchandise as a token of our appreciation!

