Daita Logo

Running Agents

Execute deployed agents and workflows remotely and manage execution history.

#run

Execute a deployed agent or workflow and stream its progress.

bash
daita run <target>

Options:

FlagDefaultDescription
--typeagentTarget type: agent or workflow
--data <file>JSON file with input data
--data-json <string>JSON string with input data (inline)
--task <name>processTask name to invoke on the agent
--follow / -fFollow progress in real-time
--timeout <seconds>300Maximum wait time

Examples:

bash
# Run an agent with inline data
daita run my_agent --data-json '{"query": "summarize Q1 sales"}'
 
# Run a workflow from a JSON file, following output
daita run my_workflow --type workflow --data input.json --follow
 
# Run a specific task
daita run my_agent --task classify --data-json '{"text": "hello"}'

The command polls until the execution completes and prints the result. In --output json mode the spinner is suppressed and the raw result JSON is written to stdout, suitable for scripting.

bash
daita run my_agent --data-json '{"q": "test"}' --output json | jq '.result'

#executions

Manage and inspect the history of agent and workflow executions.

#List executions

bash
daita executions list

Options:

FlagDefaultDescription
--limit10Number of executions to return
--statusFilter by status: queued, running, completed, failed, cancelled
--typeFilter by type: agent or workflow
bash
# Recent failures
daita executions list --status failed --limit 20

#Show execution details

bash
daita executions show <execution_id>

Returns the full execution record including status, input data, result, timing, and any error.

#Stream execution logs

bash
daita executions logs <execution_id>

Options:

FlagDescription
--follow / -fPoll until the execution completes, printing new log lines as they arrive
bash
daita executions logs exec_abc123 --follow

#Cancel an execution

bash
daita executions cancel <execution_id>

Sends a cancellation request for a queued or running execution.