Skip to main content

Dargus

Clinical efficacy prediction system for drug-development researchers.

Installation

Install Dargus with one terminal command (Linux + macOS). The installer bootstraps uv, provisions a uv-managed Python when needed, and installs the dargus command in an isolated tool environment — it never touches an existing conda/system Python:

curl -LsSf https://github.com/Moke110/Dargus/releases/latest/download/install.sh | sh

On Linux servers this is the same path used in WSL.

From a source checkout (developers)

pip install -e ".[dev]"

First run

Once installed, initialise Dargus on your machine:

dargus setup

The interactive wizard confirms your Dargus home (default ~/.dargus/, or $DARGUS_HOME if you set it) and sets up everything under it:

  • a clean default config (dargus_config.yaml)
  • your LLM API key in a restricted-permission {home}/.env
  • the D-Base directory structure ({home}/dbase)
  • migration of any legacy per-workspace session archives into {home}/sessions

Until setup has run, one-shot commands refuse with "run dargus setup first" and the bare REPL shows a setup banner.

Usage

Dargus provides a unified CLI.

One-shot commands

# Send a query to Iris
dargus iris "how does aspirin work?"

# Configuration menu
dargus config

# Test menu
dargus test

# Run the interactive setup wizard
dargus setup

Interactive REPL

# Launch interactive REPL
dargus

In the REPL, you can:

  • Type natural language queries directly
  • Use /help to see available commands
  • Use /config to configure LLM
  • Use /test to run the test suite
  • Use /clear-dbase to clear all records from the global D-Base
  • Use /new / /resume <id> to start or resume a session
  • Use /quit to exit

Upgrading

uv tool upgrade dargus-cli

Upgrading never touches your Dargus home data.

Uninstalling

dargus uninstall

Uninstall removes the program (via uv) and preserves your Dargus home data — config, secrets, D-Base, and session archive — printing where they remain. Nothing of yours is deleted; remove the Dargus home directory yourself if you also want to delete the data.

Where your data lives (Dargus home)

One per-user home holds everything ($DARGUS_HOME, default ~/.dargus/):

Data Location
Configuration {home}/dargus_config.yaml
Secrets (API keys) {home}/.env
D-Base evidence store {home}/dbase
Session archive {home}/sessions

Architecture

  • CLI (dargus/cli/) — User interface layer, includes both one-shot commands and REPL
  • API (dargus/api.py) — Public facade, all UI access goes through here
  • Core (dargus/iris/, dargus/dbase/, dargus/runtime/, dargus/agents/) — Internal runtime: the PRA (Perceive → Reason → Act) agent harness, Iris, and the D-Base evidence store

Development

# Run tests
pytest

# Lint
ruff check dargus

# Format
black dargus dargus/tests

Release files for dargus-cli 0.19.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for dargus-cli 0.19.0
File Size Uploaded
dargus_cli-0.19.0.tar.gz 137.7 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for dargus-cli 0.19.0
File Interpreter ABI Platform
dargus_cli-0.19.0-py3-none-any.whl Python 3 none any Details

Total release size: 317.0 kB

Release files / dargus_cli-0.19.0.tar.gz

Download URL dargus_cli-0.19.0.tar.gz
Size 137.7 kB
Tags Source
SHA-256 checksum
How to use checksums
08adb02a37197226672c2a5a2f5d5cc4c915fcbcf941919055a955fe9acaf9fc
BLAKE2b-256 checksum
How to use checksums
56e68649993987142d65c676687a7ca978c516f6b8ba441c8162e3916ec40e0d
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 12, 2026.

Transparency log

Release files / dargus_cli-0.19.0-py3-none-any.whl

Download URL dargus_cli-0.19.0-py3-none-any.whl
Size 179.3 kB
Tags Python 3
SHA-256 checksum
How to use checksums
9b09a786fa1eef4a61c611e2f240c36027ba0f907cf0720931204ff9e274cf2c
BLAKE2b-256 checksum
How to use checksums
0803197afc0a46860c211483c4e5234b6aef0d93b3cf8bf862ee87f3faaa132f
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Aug 12, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.19.0 This release

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page