Skip to content

gpt4free

v7.8.0 Breaking

This release includes breaking changes for platform teams planning a safe upgrade.

Published 17d LLM Frameworks
✓ No known CVEs patched
Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

llm chatbots chatgpt chatgpt-4 chatgpt-api chatgpt-free
+13 more
chatgpt4 deepseek deepseek-api deepseek-r1 gpt-4 gpt-4o gpt4 gpt4-api language-model openai openai-api openai-chatgpt reverse-engineering

ReleasePort's take

Light signal
editorial:auto 11d

Version v7.8.0 introduces system tray integration for GPT4Free UI and adds request‑logging middleware, while critical bug fixes address module lazy loading errors and path validation issues.

Why it matters: Bug fixes with severity 40 resolve crashes caused by lazy loading and improve API key handling; developers should test these changes before deployment.

Summary

AI summary

Broad release touches @hlohaus, Download Options Python Package:, 📚 Documentation & Project, and 🔧 Bug Fixes.

Changes in this release

Feature Low

Adds system tray integration with file and web operation tools.

Adds system tray integration with file and web operation tools.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Feature Low

Adds request logging middleware and Logs Inspector API for debugging.

Adds request logging middleware and Logs Inspector API for debugging.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Feature Low

Adds `compress` parameter to `create_app` for flexible app initialization.

Adds `compress` parameter to `create_app` for flexible app initialization.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Feature Low

Updates DeepInfra provider to implement Turnstile Captcha bypass using DrissionPage with async token retrieval.

Updates DeepInfra provider to implement Turnstile Captcha bypass using DrissionPage with async token retrieval.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Feature Low

Enhances Qwen provider with improved token retrieval and model handling.

Enhances Qwen provider with improved token retrieval and model handling.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Feature Low

Improves MCP server with multiple updates to `mcp.py` for better model context protocol integration.

Improves MCP server with multiple updates to `mcp.py` for better model context protocol integration.

Source: granite4.1:30b@2026-07-16-audit

Confidence: low

 —
Feature Low

Adds `socat` command to `start-browser.sh` for TCP forwarding support, enhancing browser startup.

Adds `socat` command to `start-browser.sh` for TCP forwarding support, enhancing browser startup.

Source: granite4.1:30b@2026-07-16-audit

Confidence: low

 —
Feature Low

Updates HuggingFace API provider implementation.

Updates HuggingFace API provider implementation.

Source: granite4.1:30b@2026-07-16-audit

Confidence: low

 —
Bugfix Medium

Fixes module lazy loading to resolve string attribute errors and offline GUI crashes.

Fixes module lazy loading to resolve string attribute errors and offline GUI crashes.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Bugfix Medium

Replaces `nest_asyncio` with `nest-asyncio2` in Debian build script to fix dependencies.

Replaces `nest_asyncio` with `nest-asyncio2` in Debian build script to fix dependencies.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Bugfix Medium

Fixes path validation logic in Api class for API key handling and demo paths.

Fixes path validation logic in Api class for API key handling and demo paths.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Bugfix Medium

Adds `.py` MIME type to safe‑types whitelist for file operations.

Adds `.py` MIME type to safe‑types whitelist for file operations.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Bugfix Medium

Removes unused tests for file listing/appending and `.py` extension from safe‑types whitelist.

Removes unused tests for file listing/appending and `.py` extension from safe‑types whitelist.

Source: llm_adapter@2026-07-16

Confidence: high

 —
Bugfix Medium

Fixes tool call indexing and yield logic in Antigravity and Gemini providers for reliable streaming.

Fixes tool call indexing and yield logic in Antigravity and Gemini providers for reliable streaming.

Source: granite4.1:30b@2026-07-16-audit

Confidence: low

 —
Bugfix Low

Corrects tool call indexing issue in GeminiCLI provider.

Corrects tool call indexing issue in GeminiCLI provider.

Source: granite4.1:30b@2026-07-16-audit

Confidence: low

 —
Full changelog

🚀 GPT4Free v7.8.0 Release Notes

— Celebrating 3 Years of GPT4Free & Community Day Tomorrow! —

📅 Community Day Tomorrow – July 10, 2026!

Join us for the GPT4Free Community Day!
We're celebrating 3 years of GPT4Free and 1 year of g4f.dev with a full event schedule:

| Time (CET) | Activity |
|---|---|
| 18:00 | Welcome & Opening – Heiner (hlohaus) |
| 18:30 | New Providers Showcase – What's new in GPT4Free |
| 19:30 | AI & Open Source Talks – Guests from the AI field |
| 21:00 | Community Q&A & Open Discussion |

What to expect: AI Innovations, Open Source, New Providers, Guests from the AI field, Live Demos & Talks

🎉 Be a participant – join the conversation!
👉 Join Discord
🌐 g4f.dev

Only 1 day to go – don't miss it!

✨ What's New in v7.7.7

This release is packed with improvements across providers, API stability, developer tooling, and system integration. Here's the full breakdown:

🆕 New Features

  • System Tray Integration – Native system tray support with new file and web operation tools (@hlohaus)
  • Request Logging Middleware & Logs Inspector API – Full request logging middleware plus a new inspector endpoint for debugging API calls (@hlohaus)
  • MCP Server Enhancements – Multiple updates to mcp.py improving model context protocol integration (@hlohaus)
  • create_app now accepts compress parameter – More flexible app initialization for deployment scenarios (@hlohaus)
  • Browser startup improvements – Added socat command to start-browser.sh for TCP forwarding support (@hlohaus)

🛠 Provider Updates

  • DeepInfra – Turnstile Captcha bypass implemented using DrissionPage with async token retrieval refactoring (@kqlio67, @hlohaus)
  • Qwen – Enhanced token retrieval logic and model handling improvements (@hlohaus)
  • HuggingFace API – Updated provider implementation (@hlohaus)
  • Antigravity & Gemini providers – Fixed tool call indexing and yield logic for more reliable streaming (@hlohaus)
  • GeminiCLI – Tool call indexing fix (@hlohaus)

🔧 Bug Fixes

  • Module lazy loading fix – Intercept module access for lazy loading, resolves string attribute errors and fixes offline GUI crash [#3469] (@kqlio67)
  • Debian build script – Replaced nest_asyncio with nest-asyncio2 to fix build dependencies (@kqlio67)
  • API key requirement path checks – Fixed path validation logic in the Api class, simplified demo path handling, and refined credential passing in log fetch (@hlohaus)
  • Provider registration – Refined model handling, provider registration, and ClientFactory usage in the Api class (@hlohaus)
  • MIME type mapping – Added .py file support in safe-types whitelist, with corresponding test cleanup (@hlohaus)
  • Test cleanup – Removed unused tests for file listing/appending and .py extension from safe-types whitelist (@hlohaus)

📚 Documentation & Project

  • New logo and images in README (@hlohaus)
  • FUNDING.yml – Updated and reordered ko-fi contributor list (@hlohaus)
  • Overall API and bot enhancements – Improved model handling and general functionality (@hlohaus)

👏 Contributors

A huge thank you to everyone who made this release possible:

  • @hlohaus – Heiner (Project maintainer)
  • @kqlio67 – Multiple fixes and features
  • All community members filing issues, testing, and supporting the project

🎉 See You at Community Day!

Tomorrow – July 10, 2026 – 18:00 CET
Join us for talks, demos, and open discussion. We want to hear from YOU.

🌐 g4f.dev • 💬 Discord • 📢 Telegram

Happy coding, and see you tomorrow! 🚀

Full Changelog: https://github.com/xtekky/gpt4free/compare/v7.7.6...v7.8.0

Download Options

Python Package:

  • PyPI: pip install g4f==v7.8.0

Executables:

  • Windows x64: g4f-windows-v7.8.0-x64.zip
  • Linux x64: g4f-linux-v7.8.0-x64
  • Linux ARM64: g4f-linux-v7.8.0-arm64
  • macOS x64: g4f-macos-v7.8.0-x64
  • macOS ARM64: g4f-macos-v7.8.0-arm64

System Packages:

  • WinGet: winget install gpt4free

Docker:

  • docker pull hlohaus789/g4f:v7.8.0
  • docker pull hlohaus789/g4f:v7.8.0-slim

💻 Having trouble with the .exe from this release?
👉 Check out the Windows launcher for GPT4Free instead:
🔗 https://github.com/gpt4free/g4f.exe 🚀

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track gpt4free

Get notified when new releases ship.

Sign up free

About gpt4free

All releases →

Related context

Beta — feedback welcome: [email protected]