WTFCode - v1.0.6

What's Changed

• Add image context support and /context commands by @pavel444-byte in #21
• Make image context one-shot by @pavel444-byte in #23
• Release wtfcode v1.0.6 by @pavel444-byte in #24

Full Changelog: v1.0.5...v1.0.6

WTFCode - 1.0.5 - Global update

What's Changed

• Simplify requirements.txt for Python 3.13 and add Windows platform markers by @pavel444-byte in #13
• Replace multiline output feature with multiline input mode by @pavel444-byte in #14
• Remove sys.path hack from web entrypoint and document module-based web launch by @pavel444-byte in #15
• Auto-create ~/.wtfcode/config.yml on Linux and macOS by @pavel444-byte in #17
• Add /mcp install to configure and enable MCP servers by @pavel444-byte in #18
• Release wtfcode 1.0.5 by @pavel444-byte in #19

Full Changelog: v1.0.4...v1.0.5

WTFCode - 1.0.4 - Bug fixes

What's Changed

• feat: add Azure OpenAI provider support by @devin-ai-integration[bot] in #4
• feat: add Llama provider support via Ollama by @devin-ai-integration[bot] in #5
• fix: cross-platform compatibility and command execution confirmation by @devin-ai-integration[bot] in #6
• Enforce project-local file access, secure subprocess execution, and typing fixes by @pavel444-byte in #7
• Make CodeAssist history provider-aware and update Streamlit UI by @pavel444-byte in #8
• Improve execute_command: threaded streaming, merged stderr/stdout, and timeout handling by @pavel444-byte in #9
• Return assistant content for Web and Desktop UIs by @pavel444-byte in #10
• Make config writes explicit and split theme apply/persist by @pavel444-byte in #11
• Release 1.0.4: Streamlit history, command streaming, assistant response, and config fixes by @pavel444-byte in #12

New Contributors

• @pavel444-byte made their first contribution in #7

Full Changelog: v1.0.3...v1.0.4
Or CHANGELOG.md

v1.0.3 — Config Reload, Gemini Fix, Type Safety & Error Recovery

WTFCode v1.0.3

What's New

🔧 /config reload Now Actually Works

The /config reload command previously called init_config(), which only creates a new config file if one doesn't exist — it never refreshed the in-memory config. Now it uses a new reload_config() function that clears and repopulates the global config dict from ~/.wtfcode/config.yml, so your changes take effect immediately without restarting.

🤖 Gemini Provider Fixed

• Gemini API calls updated from the deprecated client.generate_content() to the correct client.models.generate_content() across all modes (agent, ask, commit message generation).
• Agent mode now works with Gemini — previously it just printed an error and returned. Now it returns text responses properly.

⚡ Model Validation Caching

_validate_model() used to hit the provider API on every CodeAssist instantiation, slowing down startup and mode switches. Model lists are now cached per-provider in a module-level dict, so repeated instantiations are instant.

🛡️ Error Recovery in Agent Loop

API calls in run_agent() and ask_only() are now wrapped in try/except. When an API call fails:

• The error is displayed cleanly to the user
• The broken response is not appended to conversation history, so the session remains usable
• Errors are logged via Python logging for diagnostics

🏷️ Type Hints for self.client

Added a class-level Union[openai.OpenAI, anthropic.Anthropic, genai.Client] type annotation for the client attribute on CodeAssist.

Added llama and openai azure providers support

🧹 Code Cleanup

• Extracted _extract_reasoning() and _display_reasoning() helper methods to eliminate duplicated reasoning/thinking extraction logic across run_agent() and ask_only().
• Added logging module for better error diagnostics.

📺 Multi-Line Real-Time Command Output

Shell command output now streams line-by-line in a Rich Live panel, so you can watch long-running commands execute in real time.

---

Installation

Fresh Install

uv pip install git+https://github.com/pavel444-byte/WTFCode.git

Upgrade from Previous Version

uv pip install --upgrade git+https://github.com/pavel444-byte/WTFCode.git

From Source

git clone https://github.com/pavel444-byte/WTFCode.git
cd WTFCode
uv sync

---

Usage

# Start the CLI
wtfcode

# Or run directly
python main.py

CLI Commands

Command	Description
/mode	Switch between Agent and Ask modes
/theme	Change terminal theme (dark, light, matrix, dracula)
/models	List and select available models
/config reload	Reload config from ~/.wtfcode/config.yml
/commit	AI-generated commit message + commit
/add {file}	Add file content to conversation context
/web	Launch Streamlit web interface
/multiline	Toggle real-time command output
/help	Show all commands
/exit	Exit WTFCode

Supported Providers

• OpenAI (default) — GPT-4o, GPT-4, etc.
• Anthropic — Claude 3.5 Sonnet, Haiku, Opus
• Google Gemini — Gemini 1.5 Flash, Pro
• OpenRouter — Access 100+ models

Configuration

Create a .env file or edit ~/.wtfcode/config.yml:

OPENAI_API_KEY=your_key_here
PROVIDER=openai
MODEL=gpt-4o

---

Requirements

• Python >= 3.13
• uv package manager

Full Changelog: v1.0.2...v1.0.3

1.0.2 Update

Release 1.0.2

Changelog:
Fixed errors on start Missing depencies
Added git connect to ai.
Added auto initilization to config
Added config file (in C:\Users%USERPROFILE%.wtfcode)
Added notifications (windows only)

1.0.1 Release

Update!!!
Changelog:
Added config file into your folder generation into C:/Users/%USER%/.wtfcode/config.yml
Added Commit command
Added Git tools for AI
Added /init Command for intilization AGENTS.md or config (if not created auto).
Thank you For using WTFCode!
Made With love ❤️

1.0.0 Release

The start of WTFCode.
Amazing code - And without 0 dollars in your provider.
Start coding with it.
Installing:
Install it via pip (pip install git+https://github.com/pavel444-byte/WTFCode)
Run it and happy with it.