Getting Started

Your first options dashboard

Guide home | Installation | Troubleshooting

This walkthrough uses supplied synthetic data. It needs no provider account or market-data request after installation. The values demonstrate the workflow. They describe no listed security.

1. Get the repository

Install Python 3.11 or later and Git first. Run these commands in a terminal:

git clone https://github.com/Iman/agent-driven-options-desk-and-skills.git
cd agent-driven-options-desk-and-skills
python3 -m venv .venv

On macOS or Linux, activate the environment:

. .venv/bin/activate

On Windows PowerShell, activate it with:

.venv\Scripts\Activate.ps1

Install the local packages:

python -m pip install -e ./engine -e ./shell
optiondesk doctor

Expected: the doctor reports an available analytics engine. A missing external provider does not prevent this file-based walkthrough.

2. Import the sample

Run all following commands from the repository root, in the same terminal.

optiondesk chain SYNTH --from-file examples/chain-synth.json --accept-data-rights --out-dir artifacts/tutorial

The sample declares its source inside the file. The acknowledgment applies to this supplied teaching sample.

Expected: the command writes artifacts/tutorial/chain_SYNTH_2026-10-08.json. The summary reports 62 contracts. Read degraded, its reason, and the source before the results. For this sample, degraded is false and provider_used reads user snapshot, because the file carries its own rate, dividend yield, and source label.

3. Calculate the desk

optiondesk greeks --out-dir artifacts/tutorial
optiondesk exposure --out-dir artifacts/tutorial
optiondesk strategy iron_condor --out-dir artifacts/tutorial
optiondesk compare --out-dir artifacts/tutorial

Expected: the directory contains the chain, Greek ladder, positioning, strategy plans, and comparison artifacts. compare builds the available structures from the saved chain. It writes a plan for every structure it can build, so the strategy command above is optional. It shows how to build one structure at a time. Two-expiry structures require a second chain, covered in Examples.

4. Open the dashboard

optiondesk dashboard --out-dir artifacts/tutorial

Open the local dashboard in the browser on the same computer. Select SYNTH and 2026-10-08.

Without the dashboard extra, the command serves the page from Python's standard library and prints a note that says so. Install -e './shell[dashboard]' for the FastAPI server. The walkthrough page is the same either way.

Read the comparison, positioning, payoff, and Greek ladder panels. Simulation and backtest panels need separate history-based calculations. A chain alone cannot populate them.

Press Ctrl+C in the terminal to stop the dashboard. Your artifacts remain in artifacts/tutorial.

5. Continue with one task

Next task	Guide
Understand the charts	Dashboard tour
Ask an agent to use these files	Agent workflows
Replace the sample with your own export	Import your data
Compare spreads or track a paper position	Examples

The sample fixes its valuation date and time to expiry for repeatable teaching. Its dates are labels for that scenario, not a claim of current data. See sample provenance before reusing its inputs.