Minimal, fast, transactional LLM coding assistant
Most mainstream coding assistants, like Cline, Cursor and Copilot, are conversational. They maintain a messaging history, on a chat interface. They are proactive, capable of reading files by themselves, making modifications, and testing those modifications.
YACT, on the other hand, is transactional. A YACT transaction is a request response pair. The request comes from the user, and contains a list of file contents, with a textual prompt. The response comes from the LLM, and either contains a list of file contents or some textual description.
Most coding agents use a set of pre-defined tools to work with your code base. LLMs use XML formated messages, to signal tool usage intent. The agent parses the response, and executes the specified tools. For example, Cline's tools are the following: https://docs.cline.bot/tools-reference/all-cline-tools
| Tool | Description |
|---|---|
write_to_file |
Create or overwrite files |
read_file |
Read file contents |
replace_in_file |
Make targeted edits to files |
search_files |
Search files using regex |
list_files |
List directory contents |
list_code_definition_names |
List code definitions in files |
YACT, on the other hand, does not have any tools, but uses Markdown formated code blocks.
| Conversational | Transactional | |
|---|---|---|
| Keeps messaging history | Yes | No, most previous messages are pruned. |
| Generation method | Retrieval Augumented Generation (RAG) | Prompt Stuffing |
| LLM interface complexity | Complex | Simple |
| Model Selection | Larger Models | Works with larger a selection of models, down to 7b parameters. |
| MCP usage | Yes | No |
| API Cost | Up to $50 per day | Up to $2 per day |
| Vibe Coding Capable | Yes | No |
| File size limitations | No limitation, however, they tend to be less effective on large source files. | Editable file sizes are limited by the LLMs output token limit. This typically means a maximum of 400 LOC per file, however keeping file sizes under 200 LOC is recomended |
| Coding standards | No official recomendations. | Strict clean coding principles are recomended. |
Download the latest binary for your operating system from the GitHub releases page.
Available binaries:
- Linux:
yact-linux-amd64 - Windows:
yact-windows-amd64.exe - macOS:
yact-darwin-amd64oryact-darwin-arm64
For example, to download and install on Linux:
cd /tmp
wget https://github.com/agabor/yact/releases/download/v0.0.0/yact-linux-amd64
chmod +x yact-linux-amd64
sudo mv yact-linux-amd64 /usr/local/bin/yOn macOS:
cd /tmp
curl -L -o yact-darwin-amd64 https://github.com/agabor/yact/releases/download/v0.0.0/yact-darwin-amd64
chmod +x yact-darwin-amd64
sudo mv yact-darwin-amd64 /usr/local/bin/yOn Windows, download the yact-windows-amd64.exe file and place it in a directory that's in your PATH, or run it directly from where you download it.
Before using y, configure your Claude API key:
y config anthropic_api_key YOUR_API_KEYYou can verify your configuration at any time:
y configGenerate code files based on a prompt:
y act "create a user authentication handler"The AI will generate complete files with code blocks. By default, files are written directly to your filesystem. Use the --safe flag to add a .new suffix to generated files for review before replacing originals:
y act --safe "add logging to the user service"Generate standalone bash scripts:
y bash "create a script that backs up my database"Ask questions about your codebase or general topics:
y ask "how does the authentication flow work?"
y ask "what are the best practices for Go error handling?"Attach files to your prompts so the AI can reference them when generating code:
y read src/models/user.go
y read src/handlers/*.go
y ask "add validation to the user model"Use glob patterns to match multiple files. View your current attachments:
y listClear all attachments:
y clearBy default, y maintains a conversation history. Use this to build on previous responses:
y ask "what functions do we need?"
y act "now implement those functions" # The AI remembers the previous questionStart a fresh conversation:
y newRetrieve the last AI response:
y lastView current settings:
y configSet configuration values:
y config anthropic_api_key your_key_here
y config claude_model claude-opus-4-1-20250805Available configuration keys:
anthropic_api_key- Your Claude API key (required)claude_model- Which Claude model to use (default: claude-haiku-4-5-20251001)
You can pipe text directly to y:
cat requirements.txt | y act
echo "fix the database connection" | y bash- File patterns: Use glob patterns with
readto attach multiple related files at once - Safe mode: Always use
--safewhen generating new code to review changes first - Context: Build up context by reading relevant files before asking complex questions
- Conversation flow: Use multi-step conversations - ask for a plan first, then generate code
- Cost tracking: The tool displays token usage and estimated costs for each API call
The following diagram illustrates YACT message types as nodes YACT commands as edges:
graph TD
File -->|ask| Question
File -->|act| Command
File -->|plan| Objective
Question -->|assistant| Answer
Answer -->|ask| Question
Command -->|assistant| Action
Action -->|act| Command
Objective -->|assistant| Plan
Plan -->|act| Command
Display command reference:
y help
y --help
y -hConfiguration and conversation history are stored in ~/.yact/:
config- Your API key and model settingscontext.json- Conversation historyattachments.json- List of attached files
"Claude API key not configured"
- Set your API key:
y config anthropic_api_key YOUR_KEY
"No files found matching pattern"
- Check that the glob pattern matches existing files
- Use exact paths if glob patterns don't work
"File not found in attachments"
- Verify the file path is correct and the file exists
- Use
y listto see currently attached files
API errors
- Verify your API key is valid
- Check your internet connection
- Ensure your Claude API account has available credits