# ikhsan3adi/gemini-web2api architecture

> Architecture diagram of the GitHub repository ikhsan3adi/gemini-web2api, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-01). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/ikhsan3adi/gemini-web2api
- Repository: https://github.com/ikhsan3adi/gemini-web2api

## Overview

gemini-web2api is a Go proxy that presents Gemini’s web interface through OpenAI-compatible and Google-native APIs. The main workflow accepts a client request, normalizes its model and prompt (including tools and images), uploads images when needed, calls Gemini, and returns a compatible response or stream. The graph preserves the distinct API adapters, multimodal support, and Gemini transport layer. Some internal call paths are unsampled; those links are documented by the README rather than inferred from package layout.

Main entry points: [`internal/gemini/client.go`](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/gemini/client.go) (Gemini Client), [`internal/server/server.go`](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/server.go) (HTTP App), [`internal/server/chat.go`](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/chat.go) (Chat Completions), [`internal/server/responses.go`](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/responses.go) (Responses API)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_runtime["Runtime Setup"]
  node_main["Process Startup<br/>[main.go]"]
  node_config["Configuration<br/>[config.go]"]
  node_app["HTTP App<br/>[server.go]"]
end

subgraph group_api["API Adapters"]
  node_chat["Chat Completions<br/>[chat.go]"]
  node_responses["Responses API<br/>[responses.go]"]
  node_googleapi["Google API<br/>[google.go]"]
end

subgraph group_conversion["Request Conversion"]
  node_catalog["Model Catalog<br/>[models.go]"]
  node_openai_format["OpenAI Conversion<br/>[openai.go]"]
  node_google_format["Google Conversion<br/>[google.go]"]
  node_responses_format["Responses Conversion<br/>[responses.go]"]
end

subgraph group_media["Multimodal Support"]
  node_image_input["Image Parsing<br/>[openai.go]"]
  node_image_tokens["Page Tokens<br/>[tokens.go]"]
  node_image_upload["Image Upload<br/>[upload.go]"]
end

subgraph group_upstream["Gemini Integration"]
  node_gemini_client["Gemini Client<br/>[client.go]"]
  node_gemini_payload["RPC Payload<br/>[payload.go]"]
  node_gemini_auth["Cookie Authentication<br/>[auth.go]"]
end

node_clients(("API Clients"))
node_gemini_web{{"Gemini Web"}}
node_image_hosts{{"Image Hosts"}}

node_clients -->|"submit requests"| node_app
node_main -->|"loads settings"| node_config
node_main -->|"starts server"| node_app
node_app -->|"routes chat"| node_chat
node_app -->|"routes responses"| node_responses
node_app -->|"routes Google API"| node_googleapi
node_chat -->|"normalizes request"| node_openai_format
node_responses -->|"converts input"| node_responses_format
node_responses_format -.->|"normalizes messages"| node_openai_format
node_googleapi -->|"converts request"| node_google_format
node_chat -->|"resolves model"| node_catalog
node_responses -->|"resolves model"| node_catalog
node_googleapi -->|"uses model catalog"| node_catalog
node_openai_format -->|"extracts images"| node_image_input
node_image_input -.->|"fetches image URLs"| node_image_hosts
node_chat -.->|"uploads attachments"| node_image_upload
node_responses -->|"uploads attachments"| node_image_upload
node_image_upload -.->|"uses upload tokens"| node_image_tokens
node_image_upload -->|"uploads image data"| node_gemini_web
node_app -->|"creates client"| node_gemini_client
node_chat -.->|"requests generation"| node_gemini_client
node_responses -->|"requests generation"| node_gemini_client
node_googleapi -.->|"requests generation"| node_gemini_client
node_gemini_client -.->|"builds RPC payload"| node_gemini_payload
node_gemini_client -->|"sends generation requests"| node_gemini_web
node_gemini_client -.->|"uses cookie auth"| node_gemini_auth
node_gemini_web -.->|"returns generated output"| node_gemini_client

click node_main "https://github.com/ikhsan3adi/gemini-web2api/blob/main/main.go"
click node_config "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/config/config.go"
click node_app "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/server.go"
click node_chat "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/chat.go"
click node_responses "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/responses.go"
click node_googleapi "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/google.go"
click node_catalog "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/models/models.go"
click node_openai_format "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/openai.go"
click node_google_format "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/google.go"
click node_responses_format "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/responses.go"
click node_image_input "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/openai.go"
click node_image_tokens "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/multimodal/tokens.go"
click node_image_upload "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/multimodal/upload.go"
click node_gemini_client "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/gemini/client.go"
click node_gemini_payload "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/gemini/payload.go"
click node_gemini_auth "https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/gemini/auth.go"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_main,node_config,node_app,node_clients toneBlue
class node_chat,node_responses,node_googleapi toneAmber
class node_catalog,node_openai_format,node_google_format,node_responses_format toneMint
class node_image_input,node_image_tokens,node_image_upload toneRose
class node_gemini_client,node_gemini_payload,node_gemini_auth,node_gemini_web,node_image_hosts toneIndigo
```

## Components

### Runtime Setup

- [Process Startup](https://github.com/ikhsan3adi/gemini-web2api/blob/main/main.go): `main.go`
- [Configuration](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/config/config.go): `internal/config/config.go`
- [HTTP App](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/server.go): `internal/server/server.go`

### API Adapters

- [Chat Completions](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/chat.go): `internal/server/chat.go`
- [Responses API](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/responses.go): `internal/server/responses.go`
- [Google API](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/server/google.go): `internal/server/google.go`

### Request Conversion

- [Model Catalog](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/models/models.go): `internal/models/models.go`
- [OpenAI Conversion](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/openai.go): `internal/format/openai.go`
- [Google Conversion](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/google.go): `internal/format/google.go`
- [Responses Conversion](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/responses.go): `internal/format/responses.go`

### Multimodal Support

- [Image Parsing](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/format/openai.go): `internal/format/openai.go`
- [Page Tokens](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/multimodal/tokens.go): `internal/multimodal/tokens.go`
- [Image Upload](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/multimodal/upload.go): `internal/multimodal/upload.go`

### Gemini Integration

- [Gemini Client](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/gemini/client.go): `internal/gemini/client.go`
- [RPC Payload](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/gemini/payload.go): `internal/gemini/payload.go`
- [Cookie Authentication](https://github.com/ikhsan3adi/gemini-web2api/blob/main/internal/gemini/auth.go): `internal/gemini/auth.go`

### Other components

- **API Clients**
- **Gemini Web**
- **Image Hosts**

## Connections

- API Clients → HTTP App: submit requests (evidence: `README.md`)
- Process Startup → Configuration: loads settings (evidence: `main.go`)
- Process Startup → HTTP App: starts server (evidence: `main.go`)
- HTTP App → Chat Completions: routes chat (evidence: `internal/server/server.go`)
- HTTP App → Responses API: routes responses (evidence: `internal/server/server.go`)
- HTTP App → Google API: routes Google API (evidence: `internal/server/server.go`)
- Chat Completions → OpenAI Conversion: normalizes request (evidence: `internal/server/chat.go`)
- Responses API → Responses Conversion: converts input (evidence: `internal/server/responses.go`)
- Responses Conversion → OpenAI Conversion: normalizes messages
- Google API → Google Conversion: converts request (evidence: `internal/server/google.go`)
- Chat Completions → Model Catalog: resolves model (evidence: `internal/server/chat.go`)
- Responses API → Model Catalog: resolves model (evidence: `internal/server/responses.go`)
- Google API → Model Catalog: uses model catalog (evidence: `internal/server/google.go`)
- OpenAI Conversion → Image Parsing: extracts images (evidence: `internal/format/openai.go`)
- Image Parsing → Image Hosts: fetches image URLs (evidence: `README.md`)
- Chat Completions → Image Upload: uploads attachments (evidence: `README.md`)
- Responses API → Image Upload: uploads attachments (evidence: `internal/server/responses.go`)
- Image Upload → Page Tokens: uses upload tokens
- Image Upload → Gemini Web: uploads image data (evidence: `internal/multimodal/upload.go`)
- HTTP App → Gemini Client: creates client (evidence: `internal/server/server.go`)
- Chat Completions → Gemini Client: requests generation (evidence: `README.md`)
- Responses API → Gemini Client: requests generation (evidence: `internal/server/responses.go`)
- Google API → Gemini Client: requests generation (evidence: `README.md`)
- Gemini Client → RPC Payload: builds RPC payload
- Gemini Client → Gemini Web: sends generation requests (evidence: `README.md`)
- Gemini Client → Cookie Authentication: uses cookie auth
- Gemini Web → Gemini Client: returns generated output
