daita-cli

Official CLI and MCP server for the Daita platform — deploy, run, and observe your hosted AI agents from the terminal or any coding agent.

---

Installation

pip install daita-cli

Requires Python 3.11+.

---

Authentication

Export your API key before running any command:

export DAITA_API_KEY=sk-...

Get your key from the Daita dashboard.

---

Quick Start

# Initialize a new project
daita init my-project

# Test locally before deploying
daita test

# Deploy to the cloud
daita push

# Run an agent remotely
daita run my-agent --data-json '{"input": "hello"}'

# Follow execution in real-time
daita run my-agent --follow

# Re-run a past execution and compare outcomes
daita replay <execution_id>
daita diff <execution_a> <execution_b>

# Diagnose setup issues
daita doctor

---

Commands

Local Development

Command	Description
daita init [name]	Scaffold a new Daita project (agents/, workflows/, skills/)
daita create agent <name>	Add a new agent from template
daita create workflow <name>	Add a new workflow from template
daita create skill <name>	Add a new skill (instructions + tools) from template
daita test [target]	Run agents/workflows locally
daita push	Deploy the current project to the cloud
daita status	Show project and deployment status

Agents & Executions

Command	Description
daita agents list	List all agents
daita agents show <id>	Show agent details
daita agents deployed	List deployed agents
daita run <target>	Execute an agent or workflow remotely
daita replay <execution_id>	Re-run an execution with identical inputs
daita diff <exec_a> <exec_b>	Compare two executions (status, duration, cost, spans)
daita executions list	List recent executions
daita executions show <id>	Show execution details and result
daita executions cancel <id>	Cancel a running execution

Diagnostics

Command	Description
daita doctor	Run environment + platform connectivity checks with copy-pasteable fixes

Observability

Command	Description
daita traces list	List execution traces
daita traces show <id>	Show trace details
daita traces spans <id>	Show span timeline (ASCII on TTY, structured JSON on pipe)
daita traces decisions <id>	Show AI decision events
daita traces stats	Trace statistics (24h/7d/30d)
daita logs	View deployment logs
daita operations list	List platform operations
daita operations stats	Operation statistics
daita memory status	Show memory system status
daita memory show <workspace>	Show workspace memory contents

Infrastructure

Command	Description
daita deployments list	List deployments
daita deployments history <project>	Deployment history
daita deployments rollback <id>	Rollback to a previous deployment
daita schedules list	List agent schedules
daita schedules pause <id>	Pause a schedule
daita schedules resume <id>	Resume a schedule
daita secrets list	List secret key names
daita secrets set <key> <value>	Store an encrypted secret
daita secrets remove <key>	Delete a secret
daita webhooks list	List webhook URLs

---

Output Formats

All commands support --output / -o:

daita agents list -o json     # JSON (default when piped)
daita agents list -o table    # ASCII table
daita agents list -o text     # Human-readable text

Output defaults to JSON automatically when stdout is not a TTY (e.g. in scripts or CI).

---

MCP Server

daita-cli ships a full Model Context Protocol server with 35 tools, letting coding agents (Claude Code, Codex, Cursor, etc.) interact with your Daita platform directly. Tools stream MCP progress notifications during long runs and return structured JSON that coding agents can act on without parsing terminal output.

Start the server

daita mcp-server

Configure in Claude Code

Add to your project's .mcp.json:

{
  "mcpServers": {
    "daita": {
      "command": "daita",
      "args": ["mcp-server"],
      "env": {
        "DAITA_API_KEY": "sk-..."
      }
    }
  }
}

Available MCP tools

Category	Tools
Agents	list_agents, get_agent, list_deployed_agents
Executions	run_agent, list_executions, get_execution, cancel_execution, get_execution_stats, replay_execution, diff_executions
Traces	list_traces, get_trace, get_trace_spans, get_trace_decisions, get_trace_stats, get_trace_timeline
Deployments	list_deployments, get_deployment_history, delete_deployment
Schedules	list_schedules, get_schedule, pause_schedule, resume_schedule
Memory	get_memory_status, get_workspace_memory
Secrets	list_secrets, set_secret, delete_secret
Webhooks	list_webhooks
Diagnostics	doctor
Local dev	init_project, create_agent, create_workflow, create_skill, test_agent

---

Environment Variables

Variable	Description
DAITA_API_KEY	API key (required)
DAITA_API_ENDPOINT	Override the API base URL (default: https://api.daita-tech.io)
DAITA_OUTPUT	Default output format: json, text, or table
DAITA_NO_SPINNER	Set to any value to disable progress spinners

---

Development

git clone https://github.com/daita-tech/daita-cli
cd daita-cli
pip install -e ".[dev]"
pytest

---

License

Apache 2.0 — see LICENSE.