Quickstart
Before you start
Section titled “Before you start”- Install Faraday: Install
- Set a model API key such as
OPENAI_API_KEY
Single Query Mode
Section titled “Single Query Mode”faraday "Summarize structure–activity relationships across KRAS inhibitors, focusing on motifs that improve binding affinity and selectivity"Faraday will run the task while logging updates to the console for real-time feedback.
Accessing the results
Section titled “Accessing the results”While the solution will be printed to the console, its often preferable to have a persistant record of the task and it’s execution.
By default, Faraday will store the results to the ./run_outputs directory.
# in the ./run_outputs directory run_20260406045042_chat_20260406045041_b524f0c9_query_20260406045041_304a95d8│ ├── agent_outputs│ │ ├── beta_lactam_5_molecules.png│ │ └── plots│ │ ├── beta_lactam_molecules_grid.png│ │ ├── binomial_vs_normal.png│ │ ├── binomial_vs_normal_data.csv│ │ ├── binomial_vs_normal_data.txt│ │ ├── binomial_vs_normal_n50_p0.5.svg│ │ └── normal_vs_binomial_n50_p0.5.png│ └── run_artifacts│ ├── events.jsonl│ ├── metadata.json│ ├── result.json│ └── trajectory.jsonSee more information about the results directory in the Results and artifacts section.
Config-based Run
Section titled “Config-based Run”The Faraday agent can be configured using a YAML file. An example config is provided in faraday.example.yaml.
To run with a config, use the --config flag:
faraday --config faraday.example.yaml "Summarize structure–activity relationships across KRAS inhibitors, focusing on motifs that improve binding affinity and selectivity"For practical examples, see YAML guidance.
Interactive Mode
Section titled “Interactive Mode”faraday