gemini-web2api

name	gemini-web2api
description	Convert Google Gemini web into an OpenAI-compatible API. Zero auth, single file, drop-in replacement.
version	1.0.0
author	Hermes Agent
license	MIT
platforms	["linux","macos","windows"]
metadata	{"hermes":{"tags":["llm","api","gemini","openai","proxy","free-tier"],"related_skills":["llama-cpp","serving-llms-vllm"]}}

gemini-web2api

Convert Google Gemini's web interface into an OpenAI-compatible API endpoint. No API keys needed — uses your Gemini web session.

Prerequisites

• Python 3.8+
• A Google account with Gemini access
• pip

Installation

pip install gemini-web2api

Quick Start

# Start the server
gemini-web2api --port 8080

# Test with curl
curl http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"gemini","messages":[{"role":"user","content":"Hello!"}]}'

Configuration

Environment Variables

• GEMINI_SESSION — Gemini web session token (auto-extracted if not set)
• PORT — Server port (default: 8080)
• HOST — Bind address (default: 0.0.0.0)

OpenAI Client Configuration

import openai

client = openai.OpenAI(
    base_url="http://localhost:8080/v1",
    api_key="not-needed"  # Any string works
)

response = client.chat.completions.create(
    model="gemini",
    messages=[{"role": "user", "content": "Explain quantum computing"}]
)

Usage Patterns

As a Drop-in OpenAI Replacement

export OPENAI_API_BASE=http://localhost:8080/v1
export OPENAI_API_KEY=not-needed

# Now any OpenAI-compatible tool works
python my_script.py

With Hermes Agent

Add to config.yaml:

providers:
  - name: gemini-free
    base_url: http://localhost:8080/v1
    api_key: not-needed
    models:
      - gemini

Streaming Support

stream = client.chat.completions.create(
    model="gemini",
    messages=[{"role": "user", "content": "Write a story"}],
    stream=True
)
for chunk in stream:
    print(chunk.choices[0].delta.content, end="")

Common Pitfalls

1. Session expiration — Gemini web sessions expire after ~2 hours. Restart the server or refresh the session token.
2. Rate limits — Gemini web has implicit rate limits. High-volume usage may trigger temporary blocks.
3. Model naming — Use model: "gemini" in requests. The server maps this to the appropriate Gemini model.
4. Tool calling — Supported but not all Gemini tools are exposed through the web interface.

Verification

# Test server is running
curl -s http://localhost:8080/v1/models | python3 -m json.tool

# Test completion
curl -s http://localhost:8080/v1/chat/completions \
  -H "Content-Type: application/json" \
  -d '{"model":"gemini","messages":[{"role":"user","content":"Say hello"}]}' \
  | python3 -m json.tool

Source

• GitHub: https://github.com/Sophomoresty/gemini-web2api
• Stars: 870 (as of 2026-06-01)
• License: MIT