Qredence Documentation

This guide runs Fleet RLM locally from source. fleet cli starts the API server and a local MLflow server for you.

Prerequisites

Python 3.11, 3.12, or 3.13
uv
An existing PostgreSQL database
A Daytona organization and Volume, plus an API key
An API key for a model provider that DSPy supports. This guide uses openai/gpt-4o-mini as an example route.

Install

Clone the repository and install the Python environment.

git clone https://github.com/Qredence/fleet-rlm.git
cd fleet-rlm
uv sync --dev

Add credentials

Copy the example environment file.

cp .env.example .env

Open .env and set these values:

Variable	Value
`FLEET_DATABASE_URL`	Your PostgreSQL connection string.
`FLEET_LLM_MODEL`	Your DSPy model route, such as `openai/gpt-4o-mini`.
`FLEET_LLM_API_KEY`	Your provider API key. Leave it empty to use the provider’s own variable, such as `OPENAI_API_KEY`, from your shell.
`FLEET_LLM_BASE_URL`	Optional. An OpenAI-compatible endpoint. It must use HTTPS, or HTTP on `localhost`.
`FLEET_DAYTONA_API_KEY`	Your Daytona API key.
`FLEET_DAYTONA_ORG_ID`	Your Daytona organization ID.
`FLEET_DAYTONA_VOLUME_ID`	Your Daytona Volume ID.

FLEET_TYPESAFE_API_KEY is optional. It enables the TypeSafe backend that preloads relevant Skills. Keep .env out of Git. To change non-secret defaults, copy config/fleet.toml.example to config/fleet.toml. See the configuration reference.

Start a Session

uv run fleet cli --session research-1

fleet cli starts the API server on 127.0.0.1:8000 and local MLflow on 127.0.0.1:5001, then opens the terminal client. If a setting is missing or a check fails, Fleet first opens the setup checklist so you can fix it. Fleet creates its tables in the existing database at startup. Server logs go to .fleet/logs/server.log.The first time you use a Daytona organization, build the sandbox snapshot. Type /setup, choose Build snapshot (Alt+B), and confirm. Wait until the snapshot is active. See Snapshots.Try a prompt such as “Use Python to calculate the first 20 Fibonacci numbers and explain the result.” Type /help to see every command.

Work with a file

For a CSV with region and revenue columns, attach it, select it as context, and ask for an artifact.

/attach ./sales.csv
/context files/sales.csv
Calculate revenue by region and save artifacts/revenue_by_region.csv.
/artifacts
/get artifacts/revenue_by_region.csv ./revenue_by_region.csv
/trace

Inspect the answer and the artifact before you rely on them.

Resume the Session later

Pass the same Session ID to return to the conversation.

uv run fleet cli --session research-1

If you omit --session, the client generates a short ID such as sess_1a2b3c4d.

To send traces to Databricks managed MLflow instead of the local server, run uv run fleet cli databricks. See the CLI reference.

Run the server without the terminal

Start only the FastAPI server when you want to call the API from your own client.

uv run fleet serve

The server listens on http://127.0.0.1:8000 and has no caller authentication. fleet serve doesn’t start MLflow. If the local MLflow server is unreachable, the server doesn’t start. Keep fleet cli running, or start MLflow separately as the CLI reference shows. See the HTTP API reference.

Fleet-RLM

Fleet Prime Agent

Fleet Pi

GEPA Omni

Qredence Skills

Fleet Reasoner

Qredence Plugins

Fleet RLM quickstart

Prerequisites

Run the server without the terminal

Next steps

​Prerequisites

​Run the server without the terminal

​Next steps

Prerequisites

Run the server without the terminal

Next steps