Using Factory AI with VibeProxy

A simplified guide for using Factory CLI (Droid) with your personal Claude and ChatGPT subscriptions through VibeProxy.

What is This?

This guide shows you how to use Factory CLI with your personal Claude Code Pro/Max and ChatGPT Plus/Pro subscriptions instead of paying for separate API access. VibeProxy acts as a bridge that handles authentication and routing automatically.

How it works:

Factory CLI  →  VibeProxy  →  [OAuth Authentication]  →  Claude / ChatGPT APIs

VibeProxy manages OAuth tokens, auto-refreshes them, routes requests, and handles API format conversion — all automatically in the background.

Prerequisites

macOS 13.0+ (Ventura or later)
Active Claude Code Pro/Max subscription for Anthropic access
Active ChatGPT Plus/Pro subscription for OpenAI Codex access
Google account for Antigravity access (provides Gemini 3.1/3 Pro models - optional)
Google Cloud account with Gemini API access for Gemini 2.x models (optional)
GitHub Copilot subscription for Copilot model access (optional) — gives Claude, GPT, and Gemini models via GitHub's API
Z.AI API key for GLM model access (optional) - get one at z.ai/manage-apikey/apikey-list
Factory CLI installed: curl -fsSL https://app.factory.ai/cli | sh

Step 1: Install VibeProxy

Download VibeProxy.app from the releases page or build from source
Install: Drag VibeProxy.app to your /Applications folder
Launch: Open VibeProxy from Applications
- If macOS blocks it: Right-click → Open, then click "Open" in the dialog

Step 2: Connect Your Accounts

Once VibeProxy is running:

Click the VibeProxy menu bar icon
Select "Open Settings"
Click "Connect" next to Claude Code
- Your browser will open for authentication
- Complete the login process
- VibeProxy will automatically detect when you're authenticated
Click "Connect" next to Codex
- Follow the same browser authentication process
- Wait for VibeProxy to confirm the connection
(Optional) Click "Connect" next to Antigravity
- Sign in with your Google account
- Grant permissions for AI model access
- This provides access to Gemini 3.1 / 3 Pro models
- VibeProxy will automatically save your credentials
(Optional) Click "Connect" next to Gemini
- Sign in with your Google account
- Select a Google Cloud project (or accept the default)
- This provides access to Gemini 2.x models
- VibeProxy will automatically save your credentials
(Optional) Click "Connect" next to GitHub Copilot
- Authenticate with your GitHub account
- Requires an active GitHub Copilot subscription
- This provides access to Claude Opus 4.6, Sonnet 4.6, Haiku 4.5, GPT-5.5, GPT-5.3-Codex, and Gemini 3.1/3 Pro via GitHub's API
(Optional) Click "Add Account" next to Z.AI GLM
- Enter your Z.AI API key (get one at z.ai/manage-apikey/apikey-list)
- This provides access to GLM-5, GLM-4.7, and other GLM models
- VibeProxy will securely store your API key

✅ The server starts automatically and runs on port 8317

Step 3: Configure Factory CLI

Edit your Factory configuration file at ~/.factory/config.json (if the file doesn't exist, create it):

{
  "custom_models": [
    {
      "model_display_name": "CC: Opus 4.7 (High)",
      "model": "claude-opus-4-7-thinking-32000",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },
    {
      "model_display_name": "CC: Opus 4.7 (Medium)",
      "model": "claude-opus-4-7-thinking-10000",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },
    {
      "model_display_name": "CC: Opus 4.7 (Low)",
      "model": "claude-opus-4-7-thinking-4000",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },
    {
      "model_display_name": "CC: Opus 4.7",
      "model": "claude-opus-4-7",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },    
    {
      "model_display_name": "CC: Sonnet 4.6",
      "model": "claude-sonnet-4-6",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },
    {
      "model_display_name": "CC: Sonnet 4.6 (Low)",
      "model": "claude-sonnet-4-6-thinking-4000",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },
    {
      "model_display_name": "CC: Sonnet 4.6 (Medium)",
      "model": "claude-sonnet-4-6-thinking-10000",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },
    {
      "model_display_name": "CC: Sonnet 4.6 (High)",
      "model": "claude-sonnet-4-6-thinking-32000",
      "base_url": "http://localhost:8317",
      "api_key": "dummy-not-used",
      "provider": "anthropic"
    },

    {
      "model_display_name": "AG: Opus 4.6 Thinking",
      "model": "gemini-claude-opus-4-6-thinking",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "AG: Sonnet 4.6 Thinking",
      "model": "gemini-claude-sonnet-4-6-thinking",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "AG: Sonnet 4.6",
      "model": "gemini-claude-sonnet-4-6",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },

    {
      "model_display_name": "GH: Opus 4.6 (via Copilot)",
      "model": "claude-opus-4.6",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GH: Sonnet 4.6 (via Copilot)",
      "model": "claude-sonnet-4.6",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GH: Haiku 4.5 (via Copilot)",
      "model": "claude-haiku-4.5",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GH: GPT-5.3-Codex (via Copilot)",
      "model": "gpt-5.3-codex",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GH: Gemini 3.1 Pro (via Copilot)",
      "model": "gemini-3.1-pro-preview",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },

    {
      "model_display_name": "GPT-5.3 Codex",
      "model": "gpt-5.3-codex",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.3 Codex (High)",
      "model": "gpt-5.3-codex(high)",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.3 Codex Spark",
      "model": "gpt-5.3-codex-spark",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.5",
      "model": "gpt-5.5",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.5 (Low)",
      "model": "gpt-5.5(low)",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.5 (High)",
      "model": "gpt-5.5(high)",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.4",
      "model": "gpt-5.4",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.4 (High)",
      "model": "gpt-5.4(high)",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.4 Mini",
      "model": "gpt-5.4-mini",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GPT-5.4 Mini (High)",
      "model": "gpt-5.4-mini(high)",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "Gemini 3.1 Pro",
      "model": "gemini-3.1-pro-preview",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "Gemini 3.1 Flash Image",
      "model": "gemini-3.1-flash-image-preview",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "Gemini 2.5 Pro",
      "model": "gemini-2.5-pro",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "Gemini 2.5 Flash",
      "model": "gemini-2.5-flash",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "Gemini 2.5 Flash Lite",
      "model": "gemini-2.5-flash-lite",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },

    {
      "model_display_name": "Qwen3 Coder Plus",
      "model": "qwen3-coder-plus",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "Qwen3 Coder Flash",
      "model": "qwen3-coder-flash",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },

    {
      "model_display_name": "GLM-5",
      "model": "glm-5",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GLM-4.7",
      "model": "glm-4.7",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GLM-4-Plus",
      "model": "glm-4-plus",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GLM-4-Air",
      "model": "glm-4-air",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    },
    {
      "model_display_name": "GLM-4-Flash",
      "model": "glm-4-flash",
      "base_url": "http://localhost:8317/v1",
      "api_key": "dummy-not-used",
      "provider": "openai"
    }
  ]
}

Step 4: Use Factory CLI

Launch Factory CLI:
```
droid
```
Select your model:
```
/model
```
Then choose from:
- claude-opus-4-7 (Claude Opus 4.7 - Most powerful)
- claude-sonnet-4-6 (Claude 4.6 Sonnet)
- claude-opus-4.6 (Claude Opus 4.6 via Copilot)
- gpt-5.5, gpt-5.4, gpt-5.3-codex, etc.
- gemini-3.1-pro-preview, gemini-3.1-flash-image-preview, gemini-2.5-pro, etc.
Start coding! Factory will now route all requests through VibeProxy, which handles authentication automatically.

Available Models

Claude Models

claude-opus-4-7 - Claude Opus 4.7 (Most powerful, latest)
claude-sonnet-4-6 - Claude 4.6 Sonnet
Extended Thinking Variants (Claude 3.7+, Opus 4/4.7, Sonnet 4/4.6):
- *-thinking-NUMBER - Custom thinking token budget (e.g., -thinking-5000)
- Recommended presets:
  - *-thinking-4000 - "Think" mode (~4K tokens)
  - *-thinking-10000 - "Think harder" mode (~10K tokens)
  - *-thinking-32000 - "Ultra think" mode (~32K tokens)

GitHub Copilot Models

If you have a GitHub Copilot subscription, VibeProxy can route requests through GitHub's API, giving you access to many of the same models via a separate quota.

Key models:

claude-opus-4.6 - Claude Opus 4.6 (3x rate)
claude-sonnet-4.6 - Claude Sonnet 4.6 (1x rate, default in Copilot)
claude-haiku-4.5 - Claude Haiku 4.5 (0.33x rate, fast)
gpt-5.3-codex - GPT-5.3 Codex (1x rate)
gemini-3.1-pro-preview - Gemini 3.1 Pro (1x rate)

Note

Copilot model names use dot-notation (claude-opus-4.6) while Claude Code subscription models use dash-notation (claude-opus-4-7). Both work through VibeProxy but route through different auth backends. Use provider: "openai" and base_url with /v1 for Copilot models.

Discovering new model names: Run copilot --help and check the --model option for the current list of available models.

Gemini Models

Claude Models via Antigravity

Antigravity provides access to Claude models with a generous usage quota (shared with Sonnet and GPT-OSS). These models are accessed via the OpenAI-compatible API format and require Antigravity authentication.

Available Models:

gemini-claude-opus-4-6-thinking - Claude Opus 4.6 with extended thinking (backend-controlled budget)
gemini-claude-sonnet-4-6-thinking - Claude Sonnet 4.6 with extended thinking (backend-controlled budget)
gemini-claude-sonnet-4-6 - Claude Sonnet 4.6 (no thinking)

Tip

See the Step 3 configuration example above for the full Factory CLI config including these models.

Gemini 3.1 / 3 Pro (via Antigravity - requires Antigravity authentication):

gemini-3.1-pro-preview - Gemini 3.1 Pro (latest preview model)
gemini-3.1-flash-image-preview - Gemini 3.1 Flash Image
gemini-3-pro-preview - Gemini 3 Pro
gemini-3-pro-image-preview - Gemini 3 Pro with enhanced vision capabilities

Gemini 2.x (via Gemini CLI - requires Gemini authentication):

gemini-2.5-pro - Gemini 2.5 Pro (Most capable production model)
gemini-2.5-flash - Gemini 2.5 Flash (Fast and efficient)
gemini-2.5-flash-lite - Gemini 2.5 Flash Lite (Lightweight and fastest)

Important

Gemini 3.1 / 3 Pro Configuration Requirements:

Authentication: Gemini 3.1 / 3 Pro models require Antigravity authentication (not Gemini CLI auth)
Provider Setting: Must use "provider": "openai" in Factory config (Antigravity uses OpenAI API format)
Available in: VibeProxy v1.0.9+ with CLIProxyAPI 6.5.1+

Connect to Antigravity in VibeProxy Settings → Click "Connect" next to Antigravity → Sign in with your Google account. After connecting, restart VibeProxy to activate Gemini 3.1 / 3 Pro access.

Qwen Models

qwen3-coder-plus - Qwen3 Coder Plus (Most capable coding model)
qwen3-coder-flash - Qwen3 Coder Flash (Fast coding assistant)

Z.AI GLM Models

glm-5 - GLM-5 (latest GLM model; Pro availability may depend on your Z.AI plan)
glm-4.7 - GLM-4.7 (stable GLM coding model)
glm-4-plus - GLM-4-Plus (Enhanced GLM model)
glm-4-air - GLM-4-Air (Balanced performance)
glm-4-flash - GLM-4-Flash (Fast and efficient)

Note

Z.AI GLM models require an API key instead of OAuth authentication. Get your API key at z.ai/manage-apikey/apikey-list and add it in VibeProxy Settings → Z.AI GLM → Add Account.

OpenAI Models

GPT-5.5 (Latest):

gpt-5.5 - Latest GPT with improved reasoning

GPT-5.4 / GPT-5.3 Codex:

gpt-5.4 - GPT-5.4 reasoning model
gpt-5.4-mini - Faster GPT-5.4 variant
gpt-5.3-codex - Codex model for coding tasks
gpt-5.3-codex-spark - Faster Codex variant for coding tasks

Reasoning Effort Control (GPT-5.x):

Use parentheses syntax to control reasoning effort:

gpt-5.5(none) - No extended reasoning
gpt-5.5(low) - Low reasoning effort
gpt-5.5(medium) - Medium reasoning effort
gpt-5.5(high) - High reasoning effort
gpt-5.5(xhigh) - Extra high reasoning effort

This works with any supported GPT-5.x model: gpt-5.5(high), gpt-5.4(medium), gpt-5.3-codex(high), etc.

No manual CLIProxyAPI update is required—VibeProxy automatically keeps CLIProxyAPI up to date via our new auto-update workflow, so you can use new models immediately.

Troubleshooting

VibeProxy Menu Bar Status

Green dot: Server is running
Red dot: Server is stopped
Click the status to toggle the server on/off

Connection Issues

Problem	Solution
Can't connect to Claude/Codex/Gemini	Re-click "Connect" in VibeProxy settings
Factory shows 404 errors	Make sure VibeProxy server is running (check menu bar)
Authentication expired	Disconnect and reconnect the service in VibeProxy
Port 8317 already in use	Quit any other instances of VibeProxy or CLIProxyAPI
Gemini returns 401 errors	Verify your Google Cloud project has Gemini API enabled

Verification Checklist

✅ VibeProxy is running (menu bar icon shows green)
✅ Services (Claude, Codex, and optionally Gemini/Copilot) show as "Connected" in settings
✅ Factory CLI config has the custom models configured
✅ droid can select your custom models
✅ Test with a simple prompt: "what day is it?"

Extended Thinking Mode

Note

The -thinking-NUMBER model naming convention is a VibeProxy-specific implementation, not an official Claude model name from Anthropic. VibeProxy intercepts these custom model names and translates them into proper API calls with the thinking parameter.

VibeProxy automatically adds extended thinking support for Claude models! Simply append a thinking suffix to any Claude model name:

Model Name Pattern: {model-name}-thinking-{NUMBER}

Recommended Presets (based on Anthropic's official guidelines):

claude-sonnet-4-6-thinking-4000 → "Think" (~4K tokens)
claude-sonnet-4-6-thinking-10000 → "Think harder" (~10K tokens)
claude-sonnet-4-6-thinking-32000 → "Ultra think" (~32K tokens)

Custom Budgets: You can specify any token budget number:

claude-sonnet-4-6-thinking-2000 → 2,000 tokens
claude-sonnet-4-6-thinking-16000 → 16,000 tokens
claude-sonnet-4-6-thinking-50000 → 50,000 tokens

How It Works:

VibeProxy's thinking proxy intercepts requests on port 8317
Recognizes the -thinking-{NUMBER} suffix
Strips the suffix from the model name
Adds the thinking parameter with the specified budget
Forwards the modified request to CLIProxyAPI

Invalid Suffix Handling: If the suffix is not a valid integer (e.g., -thinking-blabla), VibeProxy strips the suffix and uses the vanilla model without thinking.

What You'll See:

Claude's step-by-step reasoning process before the final answer
More detailed analysis for complex problems
Transparent thought process in the response

Supported Models:

Claude Opus 4.7 (claude-opus-4-7*)
Claude Sonnet 4.6 (claude-sonnet-4-6*)

This works seamlessly with Factory CLI - just select the thinking variant in your model selector!

Interleaved Thinking (Automatic)

When you use extended thinking (-thinking-* suffix), VibeProxy automatically enables interleaved thinking by adding the anthropic-beta: interleaved-thinking-2025-05-14 header to your requests.

What is Interleaved Thinking?

Without interleaved thinking, Claude thinks once at the beginning and then executes all tool calls. With interleaved thinking enabled, Claude can think between tool calls, allowing it to:

Reason about tool results before deciding what to do next
Chain multiple tool calls with reasoning steps in between
Make more nuanced decisions based on intermediate results
Adapt its approach as it learns more from each tool interaction

Why This Matters for Coding Agents

AI coding tools like Factory Droids heavily rely on tool use (reading files, searching code, running commands). Interleaved thinking significantly improves the quality of multi-step coding tasks because Claude can:

Read a file → think about what it found → decide which file to read next
Run a test → analyze the failure → make a targeted fix
Search for a pattern → reason about the results → refine the search

Automatic Enablement

You don't need to configure anything - when you use any -thinking-* model variant, VibeProxy automatically:

Adds the thinking parameter to the request body
Injects the anthropic-beta: interleaved-thinking-2025-05-14 header
Merges with any existing beta headers (no duplicates)

This is enabled by default because if you're opting into extended thinking, you almost certainly want the improved reasoning that interleaved thinking provides for tool-heavy workflows.

Tips

Launch at Login: Enable in VibeProxy settings to auto-start the server
Auth Folder: Click "Open Folder" in settings to view authentication tokens
Server Control: VibeProxy automatically stops the server and releases port 8317 when you quit

Security

All authentication tokens are stored locally in ~/.cli-proxy-api/
Token files are secured with proper permissions (0600)
VibeProxy only binds to localhost (127.0.0.1)
All upstream traffic uses HTTPS
Tokens are auto-refreshed before expiration

Warning

By using this VibeProxy, you acknowledge and accept the following:

Terms of Service Risk: This approach may violate the Terms of Service of AI model providers (Anthropic, OpenAI, etc.). You are solely responsible for ensuring compliance with all applicable terms and policies.
Account Risk: Model providers may detect this usage pattern and take punitive action, including but not limited to account suspension, permanent ban, or loss of access to paid subscriptions.
No Guarantees: Providers may change their APIs, authentication mechanisms, or policies at any time, rendering this method inoperable without notice.
Assumption of Risk: By proceeding, you assume all legal, financial, and technical risks. The authors and contributors of this guide and CLIProxyAPI bear no responsibility for any consequences arising from your use of this method.

Use at your own risk. Proceed only if you understand and accept these risks.

Acknowledgments

VibeProxy is built on top of CLIProxyAPI, an excellent unified proxy server for AI services. Without CLIProxyAPI's robust OAuth handling, token management, and API routing capabilities, this application would not be possible.

Special thanks to the CLIProxyAPI project and its contributors for creating the foundation that makes VibeProxy work.

References

CLIProxyAPI: https://github.com/router-for-me/CLIProxyAPI
Factory CLI: https://docs.factory.ai/cli
Original Setup Guide: https://gist.github.com/ben-vargas/9f1a14ac5f78d10eba56be437b7c76e5

Need Help?

Report issues: GitHub Issues
VibeProxy by Automaze, Ltd.

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Using Factory AI with VibeProxy

What is This?

Prerequisites

Step 1: Install VibeProxy

Step 2: Connect Your Accounts

Step 3: Configure Factory CLI

Step 4: Use Factory CLI

Available Models

Claude Models

GitHub Copilot Models

Gemini Models

Claude Models via Antigravity

Qwen Models

Z.AI GLM Models

OpenAI Models

Troubleshooting

VibeProxy Menu Bar Status

Connection Issues

Verification Checklist

Extended Thinking Mode

Interleaved Thinking (Automatic)

Tips

Security

Acknowledgments

References

Uh oh!

FilesExpand file tree

FACTORY_SETUP.md

Latest commit

History

FACTORY_SETUP.md

File metadata and controls

Using Factory AI with VibeProxy

What is This?

Prerequisites

Step 1: Install VibeProxy

Step 2: Connect Your Accounts

Step 3: Configure Factory CLI

Step 4: Use Factory CLI

Available Models

Claude Models

GitHub Copilot Models

Gemini Models

Claude Models via Antigravity

Qwen Models

Z.AI GLM Models

OpenAI Models

Troubleshooting

VibeProxy Menu Bar Status

Connection Issues

Verification Checklist

Extended Thinking Mode

Interleaved Thinking (Automatic)

Tips

Security

Acknowledgments

References