Skip to main content
The adk CLI is how you scaffold, run, deploy, and debug ADK agents. Run adk --help for the top-level command list, or adk <command> --help for flags on a specific command.

Global flags

These flags work with every command:
FlagDescription
--help, -hShow help information
--version, -VShow version number
--no-cacheDisable caching for integration lookups
--profile <profile>Credentials profile to use (overrides the ADK_PROFILE environment variable and the current profile)
Most commands also accept --format json for machine-readable output. The following commands don’t: dev, login, profiles list, profiles set, upgrade, remove, self-upgrade, telemetry, theme, mcp, mcp:init, run, assets pull.

Project

These commands manage an agent project from scaffold to deploy:
CommandDescription
adk init [name]Initialize a new project
adk devStart development mode with hot reloading
adk buildBuild the agent for production
adk deployDeploy the agent to Botpress Cloud
adk checkValidate project structure and config (no login required)
adk statusShow project status, integrations, and server state
adk linkLink local agent to a workspace and bot

adk init

Scaffold a new agent project. Pass a name, or omit it to be prompted:
adk init my-agent
adk init my-agent --template hello-world
adk init --list-templates
FlagDescription
-t, --template <template>Template to use (default: blank)
-y, --yes, --defaultsSkip prompts, use defaults
--skip-linkSkip the linking step
--list-templatesList available templates and exit

adk dev

Start the dev server with hot reload:
adk dev
adk dev --port 3000 --port-console 3001
FlagDescriptionDefault
-p, --port <port>Bot server port3000
--port-console <port>Dev console port3001
--otlpEnable OTLP export to external collector (port 4318)
--port-otlp <port>Override the OTLP collector endpoint port
-v, --verboseShow additional details
--non-interactiveEmit structured NDJSON events to stdout

adk deploy

Deploy the built agent to Botpress Cloud:
adk deploy
adk deploy -y
FlagDescriptionDefault
-e, --env <environment>Deployment environmentproduction
-y, --yesAuto-approve preflight changes
Link the local agent to a workspace and bot. Writes to agent.json, or to agent.local.json with --local:
adk link
adk link --workspace wkspace_123 --bot bot_456
FlagDescription
--workspace <id>Workspace ID
--bot <id>Bot ID to link to
--dev <id>Dev bot ID
--api-url <url>Botpress API URL
-f, --forceOverwrite existing agent.json
--localWrite to gitignored agent.local.json (for multi-dev workflows)

Integrations

These commands add, inspect, and manage the integrations your agent depends on:
CommandDescription
adk add <resource>Add an integration or interface (aliases: i, install)
adk remove [resource]Remove an integration, interface, or plugin (alias: rm)
adk upgrade [integration]Upgrade integration(s) to latest version (alias: up)
adk search <query>Search the Botpress Hub
adk listList installed integrations
adk info <name>Show detailed integration info

adk add

Add an integration or interface to your agent. Accepts a name, workspace/name, or a specific version:
adk add webchat
adk add slack@latest
adk add slack@0.5.0
adk add my-workspace/custom@1.0.0
adk add interface:translator@1.0.0
adk add webchat --alias custom-webchat
FlagDescription
--alias <alias>Custom alias for the resource
Search the Botpress Hub for integrations:
adk search crm
adk search slack --limit 5
FlagDescriptionDefault
--limit <number>Max results to return20

adk list

List integrations in your project, or all available ones on the Hub:
adk list
adk list --available
FlagDescriptionDefault
--availableList all available integrations (not just installed)
--limit <n>Max results50

adk info

Show details for a specific integration. Filter by a single facet or show the full spec:
adk info slack
adk info slack --actions
adk info slack --full
FlagDescription
--actionsShow only actions
--channelsShow only channels
--eventsShow only events
--fullShow all details

Configuration and secrets

These commands manage config values and secrets:
CommandDescription
adk configValidate and fill missing config values interactively
adk config:get <key>Get a configuration value
adk config:set <key> <value>Set a configuration value
adk secretShow declared secrets and their status
adk secret:set <key> <value>Set a secret value
adk secret:delete <key>Delete a secret
adk modelsList available LLM models
All configuration and secret commands accept --prod to target the production environment instead of dev.

Chat and testing

These commands let you send messages to your agent and run eval suites:
CommandDescription
adk chatInteractive chat with your agent
adk evals [name]Run eval suites
adk evals runs [runId]List or show eval run history

adk chat

Open an interactive chat, or send a single message with --single. Requires adk dev to be running:
adk chat
adk chat --single "What's the status of order 12345?"
adk chat --single "Follow up" --conversation-id <id>
FlagDescriptionDefault
--single <message>Send one message and exit
--conversation-id <id>Continue a conversation
--timeout <duration>Max wait duration60s

adk evals

Run eval suites. With no arguments, runs all evals; pass a name to run just one:
adk evals
adk evals greeting
adk evals --tag smoke
adk evals --type regression
adk evals -v
FlagDescriptionDefault
--tag <tag>Run only evals with this tag
--type <type>Run only capability or regression evals
--judge-model <model>Model for llm_judge assertions
-v, --verboseShow full details for all evals
--server <url>Dev server URLhttp://localhost:3001

adk evals runs

List past eval runs, or show the details of a specific one:
adk evals runs
adk evals runs --latest
adk evals runs <run-id> -v
FlagDescriptionDefault
--latestShow the latest run
--limit <n>Max runs to list10
-v, --verboseShow full details

Workflows

These commands discover and run workflows against the local dev server:
CommandDescription
adk workflowsList all discovered workflows
adk workflows inspect <name>Inspect a workflow schema and metadata
adk workflows run <name> [payload]Run a workflow

adk workflows run

Run a workflow with a JSON payload. Add --wait to block until it finishes:
adk workflows run onboarding '{"userId":"123"}'
adk workflows run onboarding '{"userId":"123"}' --wait --timeout 30s
FlagDescription
--waitWait for the workflow to finish
--timeout <duration>Max wait duration (implies --wait)

Debugging

These commands help you inspect what your agent is doing:
CommandDescription
adk logs [tokens...]Query dev server logs
adk traces [tokens...]Query trace data
adk run <script> [args...]Run a TypeScript script with the ADK runtime

adk logs

Query logs from .adk/logs/. Pass filter tokens as positional arguments, or stream with --follow:
adk logs
adk logs error
adk logs warning since=1h
adk logs --follow
adk logs --follow limit=10
TokenDescription
errorShow errors only
warningShow errors + warnings
infoShow errors + warnings + info
since=<duration>Only entries newer than duration (e.g. 30s, 5m, 1h)
limit=<n>Max entries to show (default: 50)
FlagDescription
-f, --followStream logs in real time
--summaryEmit a single JSON summary (requires --format json)

adk traces

Query trace data from the local store. Filter tokens narrow by workflow, action, trigger, or drill into a specific trace:
adk traces
adk traces error
adk traces since=1h
adk traces workflow=onboarding
adk traces trace=<id>
adk traces trace=<id> --include-llm
adk traces --follow
TokenDescription
errorShow only error traces
workflow=<name>Filter by workflow name
action=<name>Filter by action/tool name
trigger=<name>Filter by trigger name
conversation=<id>Filter by conversation ID
trace=<id>Drill into a specific trace
since=<duration>Only traces newer than duration
until=<duration>Only traces older than duration
limit=<n>Max traces to show (default: 20)
FlagDescription
-f, --followStream new traces
--include-llmInclude LLM instructions and code in drill-in mode

adk run

Run a TypeScript script with the full ADK runtime, so you can import actions, Zai, and tables directly:
adk run scripts/migrate.ts
adk run scripts/seed.ts --prod
adk run scripts/process.ts arg1 arg2
FlagDescription
-f, --forceForce regeneration of the bot project
--prodUse production bot (default: dev)

Knowledge bases

Sync local knowledge base content with Botpress:
CommandDescription
adk kb syncSynchronize knowledge bases with Botpress

adk kb sync

Push local KB sources to the dev or prod bot. Use --dry-run to preview changes first:
adk kb sync --dev
adk kb sync --prod
adk kb sync --prod --force
adk kb sync --dry-run
FlagDescription
--devSync with development bot
--prodSync with production bot
--dry-runPreview changes without applying
-y, --yesSkip confirmation prompts
--forceForce re-sync all knowledge bases

Assets

Manage static files your agent serves or references:
CommandDescription
adk assets syncSynchronize assets with remote storage
adk assets listList all asset files
adk assets statusShow asset sync status
adk assets pullDownload remote assets to local directory

adk assets sync

Upload local assets to remote storage. Use --dry-run to see what would change:
adk assets sync
adk assets sync --dry-run
adk assets sync --force -y
FlagDescription
--dry-runPreview changes without applying
-y, --yesSkip confirmation prompts
--bail-on-failureStop on first error
--forceForce re-upload all files

adk assets list

List asset files. By default shows both local and remote; filter with --local or --remote:
adk assets list
adk assets list --local
adk assets list --remote
FlagDescription
--localShow only local assets
--remoteShow only remote assets

AI assistants

These commands integrate the ADK with AI coding tools:
CommandDescription
adk mcpStart the MCP server
adk mcp:initGenerate MCP configuration files

adk mcp

Start the MCP server so tools like Claude Code, Cursor, or VS Code can talk to your running dev server:
adk mcp
adk mcp --port 3001
FlagDescriptionDefault
--cwd <path>Working directory for MCP operations
--port <port>UI server port to connect to3001

adk mcp:init

Generate MCP configuration files for supported AI assistants:
adk mcp:init --all
adk mcp:init --tool claude-code
adk mcp:init --tool vscode cursor
FlagDescription
--tool <tool...>Specific tool(s): claude-code, vscode, cursor
--allGenerate for all supported tools
--forceOverwrite existing config
--project-dir <path>ADK project subdirectory (for monorepos)

Account

These commands manage your Botpress credentials and the CLI itself:
CommandDescription
adk loginAuthenticate with Botpress
adk profiles listList authentication profiles
adk profiles set [profile]Switch to a different profile
adk self-upgrade [tag]Upgrade ADK CLI to latest (or specific) version (alias: self-update)
adk telemetryManage telemetry preferences
adk themeManage CLI theme (dark/light)

adk login

Authenticate with your Botpress account. Pass --token to skip the browser flow:
adk login
adk login --token <token>
adk login --profile staging
FlagDescriptionDefault
--token <token>Botpress API token
--profile <profile>Profile name to save credentials under
--api-url <url>Botpress API URLhttps://api.botpress.cloud

adk self-upgrade

Upgrade the ADK CLI. Pass a tag (beta, next) or an explicit version:
adk self-upgrade
adk self-upgrade beta
adk self-upgrade 1.18.0

adk telemetry

View or change telemetry preferences. With no flags, prints the current status:
adk telemetry
adk telemetry --enable
adk telemetry --disable
FlagDescription
--statusShow telemetry status
--enableEnable telemetry
--disableDisable telemetry

adk theme

Switch the CLI theme between dark and light:
adk theme --set dark
adk theme --set light
FlagDescription
--set <theme>Set theme (dark or light)
Last modified on April 24, 2026