AgentChanti
Multi-agent coding assistant that plans, writes, reviews and tests code via local or cloud LLMs.
Tested on Linux
Use it
Linux x86_64
You need first
- Runtimes:
python>=3.10 - System packages:
Git - Keys or accounts: cloud API key
Install agentchanti via uv tool
uv tool install agentchanti
Where the README and reality differ
- README: kb commands listed as `agentchanti kb embed` / `agentchanti kb search "query"` with no `kb index` step mentioned
What worked: run `agentchanti kb index` in the project directory before `kb embed`/`kb search`; search must also run from the indexed project directory (index is CWD-scoped)
Check that it works
agentchanti --help 2>&1The output should contain “AgentChanti — Multi-Agent Local Coder”.
First thing to try
mkdir -p ragdemo && cat > ragdemo/calc.py <<'EOF'
def add(a, b):
"""Add two numbers."""
return a + b
class Calculator:
"""Simple calculator."""
def multiply(self, a, b):
"""Multiply two numbers."""
return a * b
EOF
echo created && ls ragdemocd ragdemo && agentchanti kb index 2>&1cd ragdemo && agentchanti kb search "multiply numbers" 2>&1In your project folder, run agentchanti kb index to index the code, then use agentchanti kb search to query it.
What it does
- Find relevant code in your project with built-in RAGTested
How to use it
- Go to your project folder and run
agentchanti kb indexto index the code. - Run
agentchanti kb search "<your query>"to list matching functions, classes and methods.
- Go to your project folder and run
- Build software from a plain English taskNot tested
- Plan, write, review and test coding tasksNot tested
- Generate and run unit tests to check codeNot tested
How far pickfit got
Linux x86_64
- Installs
- Starts
- 1 of 4 features work
- Works offline
Good for
- build software from a plain English task description
- generate and run unit tests to verify code
- break a coding task into actionable steps
How pickfit tested it
Each run starts from a clean machine and follows the steps above. How pickfit verifies
Linux x86_64
Its main function works
- Replayed from scratch
- Works offline
- Checked on 30 Oct 2026
- Version 0.13.0