agentbreeder

CLI Reference

Complete reference for all agentbreeder CLI commands.

AgentBreeder's CLI is the primary interface for managing agents. All commands support --json output for scripting and CI.

TL;DR

If you remember nothing else:

agentbreeder quickstart        # full local stack + 5 sample agents
agentbreeder init              # scaffold a new agent.yaml
agentbreeder validate          # check agent.yaml
agentbreeder deploy --target   # local, ecs-fargate, cloud-run, container-apps ship full governance; app-runner is single-container; k8s + claude-managed in flight (see deploy table)
agentbreeder chat <name>       # talk to a deployed agent
agentbreeder list agents       # see what's deployed
agentbreeder studio            # open Studio
agentbreeder down              # stop the local stack

Most-used sections below: quickstart · init · deploy · chat · list · logs · secret · provider.

Global Usage

agentbreeder [COMMAND] [OPTIONS]

Commands

agentbreeder login / logout / whoami

Authenticate against the AgentBreeder API. Tokens are stored in the OS keychain (macOS Keychain, GNOME libsecret, Windows Credential Manager). The AGENTBREEDER_API_TOKEN env var always wins over the keychain, so the same binary works on laptops and in CI.

agentbreeder login [--email EMAIL] [--password PWD] [--token JWT] [--api-url URL]
agentbreeder logout
agentbreeder whoami [--json]
OptionDescription
--email, -eAccount email. Prompted if omitted.
--password, -pAccount password. Prompted (hidden) if omitted.
--tokenPaste an existing JWT (e.g. issued by Studio) instead of using email/password.
--api-urlOverride the API URL. Defaults to $AGENTBREEDER_URL, then http://localhost:8000.

Examples:

agentbreeder login                                              # Interactive
agentbreeder login -e alice@example.com -p "$AB_PASSWORD"       # Non-interactive
agentbreeder login --token "eyJhbGciOi…"                        # Paste a token from Studio

agentbreeder whoami
# email           alice@example.com
# role            deployer
# team (primary)  engineering
# active team     ops

agentbreeder logout                                             # Clears the keychain entry

logout deletes the keychain entry. It does not revoke the token server-side — the JWT remains valid until its 24h expiry. See refresh-token issue #395 for revocation support.


agentbreeder context

Set the active team context. The active team is sent on every authenticated request as the X-AgentBreeder-Team header and is shown by agentbreeder whoami.

agentbreeder context use TEAM
agentbreeder context show
agentbreeder context clear

The context lives in ~/.agentbreeder/context.json.

Examples:

agentbreeder context use ops          # Act as team "ops"
agentbreeder context show             # Active team: ops
agentbreeder context clear            # Fall back to your primary team

agentbreeder doctor

Check that this machine satisfies AgentBreeder's prerequisites. Run before quickstart (or any time a command exits with a missing-prerequisite error) to find out exactly what's missing and how to fix it.

agentbreeder doctor [--json]

Checks performed (each prints a copy-pasteable fix when it fails):

CheckRequired
Python ≥ 3.11yes
Container runtime installed and daemon reachable (Docker, Podman, or nerdctl)yes
Free disk ≥ 8 GiB in the current working directoryyes
RAM ≥ 4 GiB (best-effort; warns on platforms where total RAM can't be detected)yes when detectable

Exits 0 when every required check passes, 1 otherwise. Designed to return in well under two seconds.

Examples:

agentbreeder doctor              # human-readable report
agentbreeder doctor --json       # JSON for scripts and CI

quickstart runs the same checks at start and bails out fast if anything's missing. Use agentbreeder quickstart --skip-doctor to bypass (advanced; intended for CI sandboxes that intentionally skip the host checks).


agentbreeder quickstart

Full local platform bootstrap — Docker, services, sample data, 5 sample agents, and Studio in one command.

agentbreeder quickstart [--cloud TARGET] [--no-browser] [--skip-seed] [--dev] [--reset]
                        [--no-ollama] [--ollama-model NAME] [--yes] [--skip-doctor]
OptionDefaultDescription
--cloudAlso deploy to cloud after local setup. aws (ECS Fargate), gcp (Cloud Run), and azure (Container Apps) all ship with full governance parity. This CLI path deploys into existing infra (plus auto-provisioned RDS/Redis data backends); greenfield provisioning ships via the Studio deploy wizard. Secret auto-mirror ships for GCP + Azure; on AWS, pre-create secrets in Secrets Manager. See the deploy-target status table.
--no-browserOffDon't open Studio at http://localhost:3001
--skip-seedOffSkip ChromaDB + Neo4j seeding (fast restart)
--devOffBuild API and Studio from local source instead of pulling images
--resetOffTear down all docker volumes (postgres, redis, chromadb, neo4j) and start fresh
--no-ollamaOffCloud-only setup. Skip the Ollama install, daemon start, and model pull (saves ~1–2 min and a ~3 GB download). Equivalent to choosing Cloud at the interactive model-source prompt.
--ollama-modelgemma3Pull a different default model for Ollama (e.g. llama3.2, phi4-mini)
--yes, -yOffNon-interactive mode for CI / scripted runs. Skip every yes/no prompt and use safe defaults: model-source falls back to the legacy "Both" path (or Cloud when --no-ollama is set), port conflicts get auto-killed, Ollama install/rebind/model-pull and Studio rebuild are auto-accepted. Blocking prompts that require human input (e.g. "press Enter once the daemon is up") exit fast instead of waiting. Does not auto-install Docker — runtime must already be present. Does not invent provider keys — export OPENAI_API_KEY (or equivalent) before running.
--skip-doctorOffSkip the prerequisite preflight (advanced; CI environments — pairs naturally with --yes).

Container-runtime diagnostics: When the daemon is installed but not reachable, quickstart names the specific cause and the exact fix command — no generic "could not connect" messages.

Detected conditionCause line + fix
Stale $DOCKER_HOST"Points at a missing socket: <path>" → unset DOCKER_HOST
Socket EACCES (Linux)"Permission denied on Docker socket" → sudo usermod -aG docker $USER && newgrp docker
Socket path mismatch (Snap/Flatpak)"CLI looking elsewhere" → export DOCKER_HOST=unix://<detected-path>
Daemon down (macOS)Lists Docker Desktop / OrbStack / Colima / Podman start commands
Daemon down (Linux)sudo systemctl start docker or systemctl --user start podman.socket
CLI missing but socket reachable"Partial install detected" → distro-specific CLI install command
Rootless DockerInfo note about bind-mount volume ownership (non-blocking)

The selected runtime line also echoes how to override it: export DOCKER_HOST=unix://<path>.

What it does:

  1. Detects Docker / Podman; shows OS-specific install instructions if neither is found
  2. Asks how you want to run models: Local (Ollama, ~3 GB), Cloud (BYO API key), or Both (default). Press Enter at the prompt to take the default; pass --no-ollama to short-circuit to Cloud. Non-interactive runs keep the legacy "Both" behavior.
  3. Creates .env with secure random secrets if not present
  4. Starts the full stack via docker-compose.quickstart.yml: API · Studio · PostgreSQL · Redis · ChromaDB · Neo4j · MCP servers · LiteLLM
  5. Waits for all services with a live progress bar
  6. Seeds ChromaDB (vector store) and Neo4j (knowledge graph) with sample data
  7. Deploys 5 sample agents: assistant, rag-agent, graph-agent, search-agent, a2a-orchestrator
  8. Opens Studio at http://localhost:3001

Examples:

agentbreeder quickstart                                  # Local only, interactive
agentbreeder quickstart --no-ollama                      # Cloud-only (skip ~3 GB download)
agentbreeder quickstart --yes --no-ollama --no-browser   # Fully scripted (CI)
agentbreeder quickstart --cloud gcp                      # Local + deploy to GCP Cloud Run (full governance + secret auto-mirror)
agentbreeder quickstart --cloud aws                      # Local + deploy to AWS ECS Fargate (full governance; pre-create secrets)
agentbreeder quickstart --cloud azure                    # Local + deploy to Azure Container Apps (full governance + secret auto-mirror)
agentbreeder quickstart --skip-seed                      # Restart stack without re-seeding data
agentbreeder quickstart --dev                            # Build from source

agentbreeder setup

Interactive wizard to configure Ollama and cloud LLM provider API keys.

agentbreeder setup [--ollama-only] [--providers-only]
OptionDescription
--ollama-onlyOnly run the Ollama setup step
--providers-onlyOnly run the cloud provider key step

Step 1 (Ollama): detects if Ollama is installed and running; shows OS-specific install instructions; presents a model picker with sizes and recommended use cases; runs ollama pull for the selected model.

Step 2 (cloud providers): walks through Anthropic, OpenAI, Google AI, and OpenRouter; shows exact URL to get each API key; validates key format; makes a live API call to verify the key works; saves to .env.

Examples:

agentbreeder setup                      # Full setup (Ollama + cloud providers)
agentbreeder setup --ollama-only        # Only configure Ollama
agentbreeder setup --providers-only     # Only configure cloud provider keys

agentbreeder seed

Seed ChromaDB (vector store), Neo4j (knowledge graph), or the AgentBreeder registry tables with sample or custom data.

agentbreeder seed [--chromadb] [--neo4j] [--registry] [--docs PATH] [--collection NAME]
                  [--cypher PATH] [--examples-dir PATH] [--clear] [--list]
OptionDefaultDescription
--chromadbOffSeed ChromaDB only
--neo4jOffSeed Neo4j only
--registryOffSeed registry tables (agents, prompts, tools, MCP servers, providers, knowledge bases) from examples/seed/. Idempotent — only empty tables are populated.
--examples-dir PATH<repo>/examples/seed/Custom path to seed YAMLs (registry mode only)
--docs PATHdeploy/seed/docs/Directory of .md / .txt files to ingest into ChromaDB
--collection NAMEagentbreeder_knowledgeChromaDB collection name
--cypher PATHCustom .cypher file to run against Neo4j
--clearOffWipe existing data before seeding
--listOffShow what is currently seeded (no writes)

Without --chromadb, --neo4j, or --registry, both ChromaDB and Neo4j are seeded.

The API server also runs the registry seeder automatically on first boot when AGENTBREEDER_AUTO_SEED=true (default for AGENTBREEDER_INSTALL_MODE=dev, disabled for cloud). Use agentbreeder seed --registry to populate registry tables manually — e.g. against an already-running API.

Examples:

agentbreeder seed                                   # Seed both ChromaDB and Neo4j
agentbreeder seed --chromadb                        # Seed ChromaDB only
agentbreeder seed --neo4j                           # Seed Neo4j only
agentbreeder seed --registry                        # Populate registry tables (agents, prompts, tools, MCP, providers, KBs)
agentbreeder seed --registry --examples-dir ./seeds/ # Use a custom seed directory
agentbreeder seed --chromadb --docs ./my-docs/      # Ingest custom .md/.txt files
agentbreeder seed --neo4j --cypher ./graph.cypher   # Load custom Cypher
agentbreeder seed --list                            # Show what's currently seeded
agentbreeder seed --clear                           # Wipe and re-seed

agentbreeder init

Scaffold a new agent project with an interactive wizard.

agentbreeder init [OUTPUT_DIR] [--lang LANG] [--framework FRAMEWORK] [--type TYPE] [--json]
Argument / OptionRequiredDefaultDescription
OUTPUT_DIRNo(agent name)Directory to create
--lang, --languageNopythonOne of python, node, go
--frameworkNo(interactive)Framework key — skips the picker
--typeNoagentOne of agent, mcp-server
--jsonNofalseJSON output for CI

What it creates (Python default):

  • agent.yaml — configuration file
  • agent.py — working example agent code
  • requirements.txt — Python dependencies
  • .env.example — environment variable template
  • README.md — getting started guide

Go scaffold (--lang go --framework custom) writes main.go, main_test.go, go.mod, agent.yaml, Dockerfile, .gitignore, .env.example, README.md. The agent uses the Go SDK and satisfies Runtime Contract v1 out of the box.

The wizard prompts for:

  1. Agent name (slug-friendly, validated)
  2. Framework (Python: LangGraph, CrewAI, Claude SDK, OpenAI, ADK, Custom; Node: Vercel AI, Mastra, LangChain.js, OpenAI Agents TS, DeepAgent, Custom; Go: Custom)
  3. Cloud target (Local, AWS, GCP, Kubernetes)
  4. Team name
  5. Owner email (auto-detected from git config)

Automatically runs agentbreeder validate after scaffolding (Python only — Go and Node validation runs at deploy time).

Examples:

agentbreeder init                                          # Interactive Python wizard
agentbreeder init my-agent                                 # Create in ./my-agent/
agentbreeder init --json                                   # JSON output (CI)
agentbreeder init --lang node --framework vercel-ai my-ts  # TypeScript agent
agentbreeder init --lang go --framework custom my-go       # Go agent
agentbreeder init --type mcp-server --lang python tools    # Python MCP server

agentbreeder deploy

Deploy an agent from an agent.yaml configuration file.

agentbreeder deploy CONFIG_PATH [--target TARGET] [--json] [--remote | --local]
Argument / OptionRequiredDefaultDescription
CONFIG_PATHYesPath to agent.yaml
--target, -tNolocalDeploy target. See supported values below.
--remoteNoAutoPOST to the API server (RBAC + audit enforced). Auto-enabled when $AGENTBREEDER_URL is set.
--localNoAutoRun the deploy engine in-process. Use for dev / offline work; bypasses RBAC + audit.
--provision, -pNoOffGreenfield-provision the cloud footprint (VPC, subnets, NAT, cluster, IAM) for a fresh account before deploying, then inject the IDs into the deploy. AWS only; local mode only for now (--local --provision). The footprint is recorded in .agentbreeder/infra-state.json and removed by agentbreeder teardown.
--jsonNoOffEmit JSON to stdout (for CI/scripting).

Remote vs. local mode

agentbreeder deploy has two execution paths:

  • Remote (recommended for production) — POSTs /api/v1/deploys against the API server identified by $AGENTBREEDER_URL, then polls the job until it terminates. The bearer token comes from agentbreeder login (OS keychain) or $AGENTBREEDER_API_TOKEN. Team-scoped RBAC, audit logging, and team-scoped cloud credentials all fire on this path.
  • Local — runs the deploy engine in the CLI process. Useful for offline development and for the local Docker Compose target. It explicitly bypasses every server-side gate, so it should never be used to deploy production agents from a developer laptop.

Mode is picked by these rules, in order:

  1. --local → local mode regardless of env
  2. --remote → remote mode regardless of env
  3. otherwise: remote when $AGENTBREEDER_URL is set, else local

Supported --target values:

This is the canonical deploy-target status table for AgentBreeder — other docs link here rather than repeating it. As of v2.6.0, the same agent.yaml deploys with full governance parity — the sidecar (guardrails, cost, tracing) plus secret mirroring — to Local, AWS ECS Fargate, GCP Cloud Run, and Azure Container Apps. AWS App Runner is single-container: it deploys without a sidecar and rejects guardrails: / secrets: at validate-infra. Kubernetes and Claude-managed are in flight.

Greenfield vs. existing account. The CLI (agentbreeder deploy) deploys into existing infrastructure — a VPC, subnets, cluster, and execution role you already have (the per-cloud BYO contract) — or, on AWS, greenfield-provisions that footprint from scratch with --provision. For GCP/Azure, greenfield ships through the Studio deploy wizard (Step 3 → Provision for me); CLI --provision for those clouds is on the roadmap. Independently, both surfaces auto-provision the data tier (RDS pgvector / ElastiCache Redis) when an agent declares memory: or a knowledge base without a backend_url.

TargetStatusGreenfield provisioningSidecar (governance)Secret auto-mirror
local (Docker Compose)✅ Shipped✅ (nothing to provision)n/a (local .env)
ecs-fargate (AWS)✅ ShippedCLI --provision + Studio wizard¹🟡 Roadmap — pre-create in Secrets Manager
cloud-run (GCP)✅ Shipped✅ Studio wizard¹ · CLI BYO + auto data tier
container-apps (Azure)✅ Shipped✅ Studio wizard¹ · CLI BYO + auto data tier
app-runner (AWS)✅ Shipped (single-container)CLI BYO only — --provision/wizard provision ECS Fargate, not App Runner❌ — rejects guardrails:/secrets:❌ — not supported
kubernetes (EKS/GKE/AKS/self-hosted)🟡 In flight❌ — needs existing cluster + kubeconfig✅ (when usable)🟡 Roadmap
claude-managed (Anthropic)🟡 In flightn/a — runtime is Anthropic-managedn/an/a

¹ AWS greenfield ships from the CLI today (agentbreeder deploy --target ecs-fargate --provision --local) and from Studio → Deploys → New deploy (the /deploy-wizard flow) + the deployments job API (POST /api/v1/deployments with infra_mode: "provision"). See Deployment → Greenfield mode and Deploying from Studio. For GCP/Azure, CLI --provision is not yet shipped — use the Studio wizard, deploy into existing infra (BYO), or pre-create the footprint first.

Each --target value maps to one cloud + runtime combination:

ValueCloudRuntime
locallocaldocker-compose
cloud-rungcpcloud-run
ecs-fargateawsecs-fargate
app-runnerawsapp-runner
container-appsazurecontainer-apps
kuberneteskubernetesdeployment
claude-managedclaude-managed(n/a — no container built)

The deploy pipeline executes 8 atomic steps:

  1. Parse & validate YAML
  2. RBAC check
  3. Dependency resolution
  4. Container build (skipped for claude-managed)
  5. Infrastructure provision
  6. Deploy & health check
  7. Auto-register in registry
  8. Return endpoint URL

If any step fails, the entire deploy rolls back.

Examples:

agentbreeder deploy ./agent.yaml                              # Deploy locally
agentbreeder deploy ./agent.yaml --target local               # Same as above (full governance)
agentbreeder deploy ./agent.yaml --target cloud-run           # GCP Cloud Run (full governance + secret auto-mirror)
agentbreeder deploy ./agent.yaml --target ecs-fargate         # AWS ECS Fargate (full governance; pre-create secrets)
agentbreeder deploy ./agent.yaml --target container-apps      # Azure Container Apps (full governance + secret auto-mirror)
agentbreeder deploy ./agent.yaml --target app-runner          # AWS App Runner (single-container; no sidecar, no secrets:)
agentbreeder deploy ./agent.yaml --target claude-managed      # Anthropic Claude Managed (managed runtime)
agentbreeder deploy ./agent.yaml --json                       # JSON output

agentbreeder validate

Validate an agent.yaml, orchestration.yaml, or memory.yaml without deploying.

agentbreeder validate CONFIG_PATH [--json]
ArgumentRequiredDescription
CONFIG_PATHYesPath to a config file

File-type detection (in order):

  1. Filename starts with orchestration → orchestration schema
  2. Filename is agent.yaml/agent.yml → agent schema
  3. Filename is memory.yaml/memory.yml or starts with memory. → memory schema
  4. Content has strategy + agents keys → orchestration schema
  5. Content has backend + memory_type (and no framework/runtime) → memory schema
  6. Content has transport/command (and no framework/runtime) → MCP (skipped)
  7. Otherwise → agent schema

Checks:

  • YAML syntax
  • JSON Schema validation (all required fields, correct types, valid enums)
  • Semantic validation (name format, version format, email format)

Examples:

agentbreeder validate ./agent.yaml
agentbreeder validate ./agent.yaml --json
agentbreeder validate ./memory.yaml          # validates against memory.schema.json

agentbreeder list

List entities from the registry.

agentbreeder list [ENTITY_TYPE] [--team TEAM] [--json]
Argument / OptionRequiredDefaultDescription
ENTITY_TYPENoagentsOne of: agents, tools, models, prompts
--teamNoFilter by team name

Examples:

agentbreeder list                           # List all agents
agentbreeder list agents                    # Same as above
agentbreeder list tools                     # List MCP servers and tools
agentbreeder list models                    # List registered models
agentbreeder list prompts                   # List prompt templates
agentbreeder list agents --team platform    # Filter by team
agentbreeder list --json                    # JSON output

agentbreeder describe

Show full details for a registered agent.

agentbreeder describe NAME [--json]
ArgumentRequiredDescription
NAMEYesAgent name

Shows: name, version, team, owner, framework, model config, tools, deploy target, status, endpoint URL, and metadata.

Examples:

agentbreeder describe my-agent
agentbreeder describe my-agent --json

Search across all registered agents, tools, models, and prompts.

agentbreeder search QUERY [--json]
ArgumentRequiredDescription
QUERYYesSearch query (keyword match)

Searches names, descriptions, tags, and teams across all entity types.

Examples:

agentbreeder search "customer support"
agentbreeder search zendesk
agentbreeder search --json "langgraph"

agentbreeder scan

Scan for MCP servers and LiteLLM models, register discoveries in the registry.

agentbreeder scan [--json]

Discovers:

  • Local MCP servers (reads schemas via MCP protocol)
  • LiteLLM models (connects to LiteLLM gateway)

Automatically registers discovered tools and models in the registry.

Examples:

agentbreeder scan
agentbreeder scan --json

agentbreeder registry

Push, list, and exercise prompts, tools, and agents in the AgentBreeder registry. Each entity has a kebab-case name, is resolved at agent startup via the file-first chain documented in the runtime, and is versioned in Postgres.

agentbreeder registry [prompt|tool|agent] SUBCOMMAND [OPTIONS]

Auth required. Every registry subcommand calls /api/v1/registry/* and needs a JWT. Mint one with POST /api/v1/auth/login, then export it:

export AGENTBREEDER_API_TOKEN=eyJhbGciOi...    # JWT from auth/login
export AGENTBREEDER_API_URL=http://localhost:8000   # default if unset

Only auth/login and auth/register are open — all 247 other management routes are gated.


agentbreeder registry prompt push

Upload a Markdown prompt file to the registry. Auto-bumps the semver patch (e.g. 1.0.0 → 1.0.1) on every save unless --version is given.

agentbreeder registry prompt push FILE [--name NAME] [--version SEMVER] [--team TEAM] [--description TEXT]
Argument / OptionTypeDefaultDescription
FILEpath.md file containing the prompt body
--namestrfilename stemKebab-case registry name (e.g. microlearning-system)
--versionsemverauto-bumpExplicit version. Omit to auto-bump patch.
--teamstrcurrent user's teamTeam that owns the prompt
--descriptionstrShort summary shown in Studio

Example:

agentbreeder registry prompt push ./prompts/microlearning-system.md \
  --name microlearning-system --team learning --description "Turns a topic into a 20-40 min ebook"
# ✓ Pushed prompt microlearning-system v1.0.2 (id: prm_a1b2…)

agentbreeder registry prompt list

List registered prompts.

agentbreeder registry prompt list [--team TEAM] [--json]
OptionTypeDefaultDescription
--teamstrFilter by team name
--jsonflagoffJSON output

Example:

agentbreeder registry prompt list --team learning
# NAME                       VERSION  TEAM      UPDATED
# microlearning-system       1.0.2    learning  2026-04-28T14:02Z

agentbreeder registry prompt try

Render a registered prompt against a real LLM (Google AI Studio / gemini-* currently) — the same path Studio's Test tab uses.

agentbreeder registry prompt try NAME --input "..." [--model MODEL] [--temperature FLOAT]
Argument / OptionTypeDefaultDescription
NAMEstrRegistered prompt name
--inputstrUser message to send
--modelstrgemini-2.5-flashAny gemini-* model
--temperaturefloat0.4Sampling temperature

Calls POST /api/v1/registry/prompts/{prompt_id}/render and returns {output, model, duration_ms, error}.

Example:

agentbreeder registry prompt try microlearning-system \
  --input "Pick a topic for a beginner ebook." \
  --model gemini-2.5-flash --temperature 0.4
# model: gemini-2.5-flash  duration: 1842ms
# output: How about "Spaced repetition for language learners"? It's narrow…

agentbreeder registry tool push

Register a tool in the registry. The CLI auto-detects how to dispatch the tool based on what you pass:

You pass…Stored endpointDispatcher at run time
./tools/get_time.pypython:<abs_path>Python subprocess (~50 ms)
./tools/get_time.tsnode:<abs_path>Node subprocess via npx tsx (~1.5 s cold)
engine.tools.standard.web_searchkept as-isIn-process Python import (~25 ms)
agentbreeder registry tool push MODULE_OR_FILE [--description TEXT]
Argument / OptionTypeDefaultDescription
MODULE_OR_FILEpath or dotted module.py / .ts / .js file, or dotted module path
--descriptionstrShort summary shown in Studio

Example:

agentbreeder registry tool push ./tools/get_current_time.py \
  --description "Returns ISO 8601 UTC time"
# ✓ Pushed tool get-current-time (endpoint: python:/abs/tools/get_current_time.py)

agentbreeder registry tool list

List registered tools and their dispatch endpoints.

agentbreeder registry tool list [--json]

Example:

agentbreeder registry tool list
# NAME                ENDPOINT                                        TYPE
# web-search          engine.tools.standard.web_search                in-process
# markdown-writer     engine.tools.standard.markdown_writer           in-process
# get-current-time    python:/abs/tools/get_current_time.py           subprocess
# get-utc-time        node:/abs/tools/get_utc_time.ts                 subprocess

agentbreeder registry tool run

Execute a registered tool with structured args. Calls POST /api/v1/registry/tools/{tool_id}/execute (requires deployer role).

agentbreeder registry tool run NAME [--args JSON] [--json]
Argument / OptionTypeDefaultDescription
NAMEstrRegistered tool name
--argsJSON string{}Tool arguments as JSON
--jsonflagoffPrint raw response JSON

Returns {output, stdout, stderr, exit_code, duration_ms, error}.

Example:

agentbreeder registry tool run web-search --args '{"query":"spaced repetition"}'
# exit_code: 0  duration: 612ms
# output: [{"title":"Spaced repetition - Wikipedia","url":"…"}, …]

agentbreeder registry agent push

Persist an agent.yaml in the registry via POST /api/v1/agents/from-yaml. The record stores name, version, framework, deploy config, and refs to the prompts + tools the agent uses.

agentbreeder registry agent push agent.yaml [--json]

Example:

agentbreeder registry agent push ./agent.yaml
# ✓ Registered microlearning-ebook-agent v0.1.0 (framework: google_adk)
#   prompts: microlearning-system@1.0.2
#   tools:   web-search, markdown-writer

agentbreeder registry agent list

List registered agents.

agentbreeder registry agent list [--team TEAM] [--json]
OptionTypeDefaultDescription
--teamstrFilter by team name

Example:

agentbreeder registry agent list --team learning
# NAME                          VERSION  FRAMEWORK   DEPLOY
# microlearning-ebook-agent     0.1.0    google_adk  cloud-run / us-central1

agentbreeder registry agent invoke

Send a message to a deployed agent's /invoke endpoint via the server-side proxy (POST /api/v1/agents/{id}/invoke). The proxy attaches the bearer token from --token (or env var AGENT_<UPPER_SNAKE>_TOKEN) so secrets never leave the API host.

agentbreeder registry agent invoke NAME --input "..." [--endpoint URL] [--token TOKEN] [--session SID] [--json]
Argument / OptionTypeDefaultDescription
NAMEstrRegistered agent name
--inputstrUser message
--endpointURLregistry valueOverride the agent's runtime URL
--tokenstrenv AGENT_<NAME>_TOKENBearer for the runtime /invoke route
--sessionstrContinue an existing session
--jsonflagoffRaw response JSON

Returns {output, session_id, duration_ms, status_code, error}.

Example:

agentbreeder registry agent invoke microlearning-ebook-agent \
  --input "What is your job?" \
  --endpoint https://microlearning-ebook-agent-sizukgalta-uc.a.run.app \
  --token $AGENT_AUTH_TOKEN
# status: 200  duration: 1.4s  session: 7c3f…
# output: My job is to turn any topic the user supplies into a polished microlearning
#         ebook a learner can complete in 20-40 minutes.

agentbreeder registry memory push

Register a memory configuration in the registry from a memory.yaml file. Calls POST /api/v1/memory/configs.

agentbreeder registry memory push FILE [--json]
Argument / OptionTypeDefaultDescription
FILEpathPath to the memory.yaml file
--jsonflagoffPrint the raw API response as JSON

memory.yaml fields used:

YAML fieldMaps to APINotes
namenameRequired, kebab-case
teamteamDefaults to default
ownerownerEmail of the responsible engineer
backendbackend_typepostgresql or redis
memory_typememory_typebuffer_window, summary, etc.
config.window_sizemax_messagesDefaults to 100
config.namespace_patternnamespace_patternOptional
config.scopescopeOptional
config.linked_agentslinked_agentsOptional list
tagstagsOptional list
descriptiondescriptionOptional

Example:

agentbreeder registry memory push ./memory.yaml
# ✔ Memory config registered: my-vercel-agent-memory  (id=1bea76e6-…)

agentbreeder registry memory list

List memory configurations from the registry. Calls GET /api/v1/memory/configs.

agentbreeder registry memory list [--json]

Example:

agentbreeder registry memory list
# NAME                          BACKEND      TYPE             MAX_MSGS
# my-vercel-agent-memory        postgresql   buffer_window    20

agentbreeder registry rag push

Register a RAG (knowledge base) index in the registry from a rag.yaml file. Calls POST /api/v1/rag/indexes.

agentbreeder registry rag push FILE [--json]
Argument / OptionTypeDefaultDescription
FILEpathPath to the rag.yaml file
--jsonflagoffPrint the raw API response as JSON

rag.yaml fields used:

YAML fieldMaps to APINotes
namenameRequired, kebab-case
descriptiondescriptionOptional
backendbackendin_memory, pgvector, or neo4j
embedding_model.provider + .nameembedding_modelJoined as "<provider>/<name>"
chunking.strategychunk_strategyDefaults to fixed_size
chunking.chunk_sizechunk_sizeDefaults to 512
chunking.chunk_overlapchunk_overlapDefaults to 64
configconfigBackend-specific config block
sourcesourceOptional source identifier
index_typeindex_typeDefaults to vector

Example:

agentbreeder registry rag push ./rag.yaml
# ✔ RAG index registered: agentbreeder-knowledge  (id=8439cd66-…)

agentbreeder registry rag list

List RAG indexes from the registry. Calls GET /api/v1/rag/indexes.

agentbreeder registry rag list [--json]

Example:

agentbreeder registry rag list
# NAME                       BACKEND     EMBEDDING                       CHUNKS
# agentbreeder-knowledge     in_memory   openai/text-embedding-3-small   512/64

agentbreeder registry rag ingest

Upload and ingest one or more files into a registered RAG index. Resolves the index name (or UUID) and POSTs multipart/form-data to /api/v1/rag/indexes/{id}/ingest. Files are chunked, embedded, and indexed by the API.

agentbreeder registry rag ingest NAME_OR_ID FILE [FILE...] [--replace] [--json]
Argument / OptionTypeDefaultDescription
NAME_OR_IDstrRegistered RAG index name (e.g. agentbreeder-knowledge) or its UUID
FILE...pathOne or more files to ingest
--replaceflagoffDelete any existing chunks whose source matches one of the incoming filenames before ingest. Without this flag, ingest is idempotent: chunks whose SHA-256 content hash already exists in the index are skipped.
--jsonflagoffPrint the raw API response as JSON

Allowed file types: .pdf, .txt, .md, .csv, .json (matches the API allow-list — unsupported extensions are rejected client-side before any upload).

Idempotency: by default, re-ingesting the same file is a no-op — the API hashes every chunk (SHA-256) and skips any chunk whose hash already exists in the index. Use --replace when you've edited the file and want the old chunks to be dropped first.

Example:

agentbreeder registry rag ingest agentbreeder-knowledge ./docs/intro.md ./docs/quickstart.pdf
# ✓ Ingested 14 chunks from 2 file(s) into agentbreeder-knowledge  (job=ca2e0c7a  status=completed)

# After editing intro.md, drop the old chunks and re-ingest:
agentbreeder registry rag ingest agentbreeder-knowledge ./docs/intro.md --replace

Run a hybrid vector + text search against a RAG index. Calls POST /api/v1/rag/search.

agentbreeder registry rag search NAME_OR_ID --query TEXT [--top-k N] [--json]
Argument / OptionTypeDefaultDescription
NAME_OR_IDstrRegistered RAG index name or UUID
--query, -qstrSearch query text
--top-k, -kint5Number of results to return
--jsonflagoffPrint the raw API response as JSON

Example:

agentbreeder registry rag search agentbreeder-knowledge -q "how do I deploy an agent?" -k 3
#  # | Score | Source            | Snippet
# ---+-------+-------------------+-----------------------------------------------
#  1 | 0.913 | docs/intro.md     | AgentBreeder compiles an agent.yaml file to…
#  2 | 0.812 | docs/quickstart.pdf | Run `agentbreeder deploy --target gcp` to…

agentbreeder logs

Show logs from a deployed agent.

agentbreeder logs AGENT_NAME [--lines N] [--follow] [--since DURATION] [--json]
Argument / OptionRequiredDefaultDescription
AGENT_NAMEYesName of the deployed agent
--lines, -nNo50Number of recent lines to show
--follow, -fNofalseStream logs in real time
--sinceNoShow logs since duration (e.g., 5m, 1h, 2d)

Examples:

agentbreeder logs my-agent                  # Last 50 lines
agentbreeder logs my-agent -n 100           # Last 100 lines
agentbreeder logs my-agent --follow         # Stream logs
agentbreeder logs my-agent --since 5m       # Logs from last 5 minutes
agentbreeder logs my-agent -f --since 1h    # Stream from last hour

agentbreeder status

Show deploy status of agents.

agentbreeder status [AGENT_NAME] [--json]
ArgumentRequiredDescription
AGENT_NAMENoAgent name (omit for all agents summary)

Without an agent name, shows a summary table of all deployed agents with their status. With an agent name, shows detailed status including deploy pipeline progress.

Examples:

agentbreeder status                  # All agents summary
agentbreeder status my-agent         # Detailed status for one agent
agentbreeder status --json           # JSON output

agentbreeder teardown

Remove a deployed agent and clean up its resources, or bulk-clean an entire cloud environment.

Single-agent teardown (local or cloud):

agentbreeder teardown AGENT_NAME [--force] [--json]
Argument / OptionRequiredDefaultDescription
AGENT_NAMEYesName of the agent to remove
--force, -fNofalseSkip confirmation prompt

Org-wide cloud teardown:

agentbreeder teardown --cloud <gcp|aws|azure|all> [options]
OptionDefaultDescription
--cloudCloud provider to clean up: gcp, aws, azure, or all
--dry-runfalsePreview what would be deleted without making any changes
--agentLimit teardown to a single named agent across cloud targets
--regionCloud region (GCP / AWS)
--projectGCP project ID
--destroy-resource-groupfalseAlso delete the Azure resource group itself

Removes Cloud Run services, ECS services/task definitions, Container Apps, Artifact Registry / ECR repositories, and associated secrets for all agents labelled managed-by: agentbreeder.

Examples:

# Single-agent local teardown
agentbreeder teardown my-agent              # Prompts for confirmation
agentbreeder teardown my-agent --force      # No confirmation

# Bulk cloud teardown
agentbreeder teardown --cloud gcp --project my-proj --region us-central1 --dry-run
agentbreeder teardown --cloud aws --region us-east-1
agentbreeder teardown --cloud azure --destroy-resource-group
agentbreeder teardown --cloud all           # All three clouds at once

# Limit to one agent across all clouds
agentbreeder teardown --cloud all --agent customer-support-agent --dry-run

agentbreeder provider

Manage LLM provider connections and API keys.

agentbreeder provider [SUBCOMMAND] [OPTIONS]

Subcommands

SubcommandDescription
listList configured providers + the OpenAI-compatible catalog
add NAMEAdd a provider (legacy types or --type openai_compatible)
test NAMETest a provider connection (catalog or legacy)
models NAMEList available models from a provider
remove NAMERemove a provider and its API key
disable NAMEDisable a provider without removing
enable NAMERe-enable a disabled provider
publish NAMEPromote a user-local entry to a PR against the upstream catalog

Hand-written providers: openai, anthropic, google, ollama, litellm, openrouter (interactive wizard).

Catalog providers (OpenAI-compatible): nvidia, openrouter, moonshot, groq, together, fireworks, deepinfra, cerebras, hyperbolic. See Providers for the full list and how to add your own.

API keys are stored in the project .env file (or ~/.agentbreeder/.env if no project). Catalog provider keys live only in your environment — only the env-var name is persisted, never the value.

Examples:

# Legacy (hand-written) providers
agentbreeder provider list                                              # List all providers + catalog
agentbreeder provider add openai                                        # Interactive setup
agentbreeder provider add openai --api-key sk-proj-...                  # Non-interactive
agentbreeder provider add ollama                                        # Auto-detect local Ollama
agentbreeder provider test openai                                       # Verify connection + latency

# OpenAI-compatible catalog (Track F)
agentbreeder provider test nvidia                                       # Built-in preset; expects NVIDIA_API_KEY
agentbreeder provider add my-vllm \
  --type openai_compatible \
  --base-url https://vllm.internal/v1 \
  --api-key-env COMPANY_VLLM_KEY                                        # User-local entry
agentbreeder provider remove my-vllm                                    # Remove user-local entry
agentbreeder provider publish my-vllm                                   # Print PR patch for upstream catalog

agentbreeder provider list --json                                       # JSON output

agentbreeder model

Manage the model registry — list, inspect, sync from configured providers, or manually deprecate a model.

agentbreeder model [SUBCOMMAND] [OPTIONS]

Subcommands

SubcommandDescription
listList models in the registry. Supports --provider, --status, --per-page, --json.
show NAMEShow a single model's full record (id, status, context, pricing, lifecycle timestamps).
syncDiscover models from each configured provider and reconcile the registry. Supports --provider NAME (repeatable). Deployer role required.
sync-nowRun the daily sync sweep immediately, in-process (bypasses the API server). Talks directly to the database; useful for testing the cron and for self-hosted environments without the daily loop. Supports --json.
deprecate NAMEManually mark a model deprecated. Supports --replacement NAME. Deployer role required.

The sync and deprecate subcommands talk to POST /api/v1/models/sync and POST /api/v1/models/{name}/deprecate respectively. Status flips emit audit events: model.added, model.deprecated, model.retired. See Providers → Model lifecycle for the full state machine.

Examples:

# Inspect
agentbreeder model list                                # All models, status badges
agentbreeder model list --provider nvidia              # Restrict to one provider
agentbreeder model list --status deprecated --json     # Machine-readable + filtered
agentbreeder model show gpt-4o                         # Full record

# Sync
agentbreeder model sync                                # Every provider with an api-key
agentbreeder model sync --provider nvidia --provider groq
agentbreeder model sync --json                         # Machine-readable result

# Sync immediately, in-process (no API server required)
agentbreeder model sync-now                            # Same sweep as the daily cron
agentbreeder model sync-now --json                     # Machine-readable summary

# Manual deprecate (operator override)
agentbreeder model deprecate gpt-3.5-turbo --replacement gpt-4o-mini

Daily cron. Cloud workspaces (AGENTBREEDER_INSTALL_MODE=cloud) run a built-in daily sweep that auto-deprecates absent models and emits a model.sync.scheduled audit event. Self-hosted users opt in via AGENTBREEDER_MODELS_DAILY_SYNC=true. Models that disappear from a provider's /models endpoint are auto-deprecated on the next sweep; after 30 days of continuous absence they're auto-retired.


agentbreeder chat

Interactive chat with a deployed agent in the terminal. Use --local to bypass the API server entirely and talk directly to a local Ollama model.

agentbreeder chat AGENT_NAME [--local] [--model MODEL] [--env ENV] [--verbose] [--json]
Argument / OptionRequiredDefaultDescription
AGENT_NAMEYesName of the agent to chat with
--local, -lNofalseBypass the API server — chat directly via Ollama (no server required)
--model, -mNoOverride the agent's configured model (or select an Ollama model in --local mode)
--env, -eNodevEnvironment (dev, staging, production)
--verbose, -vNofalseShow tool calls, token counts, latency

In-chat commands: /help, /clear, /quit (or /exit, /q)

On exit, displays a session summary with turn count, total tokens, and cost.

JSON mode (--json) reads messages from stdin (one per line) and writes JSON responses to stdout — useful for scripting and CI.

Examples:

agentbreeder chat my-agent                          # Interactive chat (API server required)
agentbreeder chat my-agent --local                  # Chat via Ollama — no API server needed
agentbreeder chat my-agent --local --model llama3.2 # Use a specific Ollama model
agentbreeder chat my-agent --verbose                # Show tool calls + costs
agentbreeder chat my-agent --model gpt-4o           # Override model
echo "hello" | agentbreeder chat my-agent --json    # JSON stdin/stdout

agentbreeder eject

Generate a Full Code SDK scaffold from an existing agent.yaml or orchestration.yaml file. Enables tier mobility from Low Code (YAML) to Full Code (Python/TypeScript SDK) without losing any configuration.

agentbreeder eject CONFIG_PATH [--sdk SDK] [--output PATH]
Argument / OptionRequiredDefaultDescription
CONFIG_PATHYesPath to agent.yaml or orchestration.yaml
--sdkNopythonTarget SDK language: python or typescript
--output, -oNoautoOutput file path

Agent ejection recreates the YAML as a builder-pattern Agent(...) chain using Agent, Tool, Model, and Memory from agenthub. Includes commented scaffolding for middleware, event hooks, and custom routing.

Orchestration ejection recreates orchestration.yaml as the appropriate SDK class (Orchestration, Pipeline, FanOut, or Supervisor) with all agents, routes, shared state, and deploy config preserved.

What it generates (agents):

  • Agent(...) builder chain matching every field in the YAML
  • Model, prompt, tools, memory, guardrails, deploy, and tags
  • Commented middleware and hook examples
  • __main__ block that validates and prints YAML round-trip

What it generates (orchestrations):

  • Correct subclass (Pipeline, FanOut, Supervisor, or Orchestration)
  • All agents, routing rules, fallbacks, shared state, supervisor config
  • Commented custom Router subclass scaffold for advanced routing logic

Examples:

# Agent ejection
agentbreeder eject agent.yaml                          # Default: agents/<name>/agent_sdk.py
agentbreeder eject agent.yaml --sdk python             # Explicit Python SDK
agentbreeder eject agent.yaml --sdk typescript         # TypeScript SDK
agentbreeder eject agent.yaml -o src/my_agent.py       # Custom output path

# Orchestration ejection
agentbreeder eject orchestration.yaml                  # Default: orchestration_sdk.py
agentbreeder eject orchestration.yaml --sdk typescript # TypeScript SDK

agentbreeder orchestration

Manage multi-agent orchestration pipelines defined in orchestration.yaml or built with the Full Code Orchestration SDK.

agentbreeder orchestration SUBCOMMAND [OPTIONS]

agentbreeder orchestration validate

Validate an orchestration.yaml file against the JSON Schema.

agentbreeder orchestration validate PATH [--json]
agentbreeder orchestration validate orchestration.yaml
agentbreeder orchestration validate pipelines/support.yaml --json

agentbreeder orchestration deploy

Validate, register, and deploy an orchestration.

agentbreeder orchestration deploy PATH [--json]
agentbreeder orchestration deploy orchestration.yaml
agentbreeder orchestration deploy pipelines/research.yaml --json

agentbreeder orchestration list

List registered orchestrations.

agentbreeder orchestration list [--team TEAM] [--status STATUS] [--json]
OptionDescription
--teamFilter by team name
--statusFilter by status: deployed, draft, error
agentbreeder orchestration list
agentbreeder orchestration list --team eng --json

agentbreeder orchestration status

Show orchestration detail and agent graph.

agentbreeder orchestration status NAME [--json]
agentbreeder orchestration status support-pipeline

agentbreeder orchestration chat

Send messages interactively to a deployed orchestration.

agentbreeder orchestration chat NAME [--verbose] [--json]
OptionDescription
--verboseShow per-agent trace, token counts, latency
--jsonRead from stdin, write JSON per line (for CI)
agentbreeder orchestration chat support-pipeline
agentbreeder orchestration chat support-pipeline --verbose
echo '{"message": "billing question"}' | agentbreeder orchestration chat support-pipeline --json

agentbreeder submit

Submit a resource for review by creating a pull request.

agentbreeder submit RESOURCE_TYPE NAME [--message MSG] [--json]
Argument / OptionRequiredDescription
RESOURCE_TYPEYesOne of: agent, prompt, tool, rag, memory
NAMEYesResource name
--message, -mNoPR description

Creates a PR from the draft branch to main. Shows PR ID, status, and diff summary.

Examples:

agentbreeder submit agent my-agent
agentbreeder submit agent my-agent -m "Added Zendesk tool integration"
agentbreeder submit prompt support-v3 --json

agentbreeder review

Review pull requests for resources.

agentbreeder review [SUBCOMMAND] [OPTIONS]

Subcommands

SubcommandDescription
listList pending reviews (filterable by --status, --type)
show PR_IDShow PR detail: diff, commits, comments
approve PR_IDApprove a PR
reject PR_ID -m "reason"Reject a PR (message required)
comment PR_ID -m "text"Add a comment to a PR

Examples:

agentbreeder review list                           # Pending reviews
agentbreeder review list --status approved         # Filter by status
agentbreeder review list --type agent              # Filter by resource type
agentbreeder review show pr-abc123                 # Show PR detail
agentbreeder review approve pr-abc123              # Approve
agentbreeder review reject pr-abc123 -m "Needs error handling"
agentbreeder review comment pr-abc123 -m "LGTM"
agentbreeder review list --json                    # JSON output

agentbreeder publish

Merge an approved PR and publish the resource to the registry.

agentbreeder publish RESOURCE_TYPE NAME [--version VERSION] [--json]
Argument / OptionRequiredDescription
RESOURCE_TYPEYesOne of: agent, prompt, tool, rag, memory
NAMEYesResource name
--version, -vNoExplicit semver tag (e.g., 2.0.0)

Finds the approved PR for the resource, merges to main, tags with semver, and publishes to the registry.

Examples:

agentbreeder publish agent my-agent                    # Auto-version
agentbreeder publish agent my-agent --version 2.0.0    # Explicit version
agentbreeder publish prompt support-v3 --json

agentbreeder secret

Manage secrets across pluggable backends (env file, AWS Secrets Manager, GCP Secret Manager, HashiCorp Vault).

agentbreeder secret [SUBCOMMAND] [OPTIONS]

agentbreeder secret list

agentbreeder secret list [--backend BACKEND] [--prefix PREFIX] [--json]
OptionDefaultDescription
--backend, -benvBackend: env, aws, gcp, vault
--prefixagentbreeder/Secret prefix (AWS/GCP/Vault)
--jsonOffOutput as JSON

Lists secret names and masked values (actual values are never printed).

agentbreeder secret set

agentbreeder secret set NAME [--value VALUE] [--backend BACKEND] [--prefix PREFIX] [--tag key=value] [--json]
Argument / OptionRequiredDescription
NAMEYesSecret name (e.g., OPENAI_API_KEY)
--value, -vNoValue (prompted securely if omitted)
--backend, -bNoBackend (env, aws, gcp, vault)
--tag, -tNokey=value tags (cloud backends only, repeatable)

agentbreeder secret get

agentbreeder secret get NAME [--backend BACKEND] [--reveal] [--json]

Prints masked value by default. Use --reveal to print the actual value.

agentbreeder secret delete

agentbreeder secret delete NAME [--backend BACKEND] [--force] [--json]

Prompts for confirmation unless --force is passed.

agentbreeder secret rotate

agentbreeder secret rotate NAME [--value NEW_VALUE] [--backend BACKEND] [--json]

Prompts for the new value with confirmation if --value is omitted. Emits a secret.rotated audit event.

agentbreeder secret sync

agentbreeder secret sync --target {aws,gcp,vault} [--workspace NAME] [--include KEY] [--exclude KEY] [--prefix PREFIX] [--dry-run] [--json]

Mirrors workspace secrets to a cloud secrets manager. This is the same operation that runs implicitly inside agentbreeder deploy when the agent declares deploy.secrets and targets aws or gcp. Run it manually to pre-populate a new environment. Each mirrored secret emits a secret.mirrored audit event.

OptionRequiredDescription
--targetYesTarget cloud: aws, gcp, or vault
--workspaceNoWorkspace name (default: workspace config)
--includeNoRestrict to specific secret names (repeatable)
--excludeNoSkip these secret names (repeatable)
--prefixNoPrefix in target backend (default: agentbreeder/)
--dry-runNoPreview without writing
--jsonNoOutput as JSON

See the Secrets guide for full workspace + auto-mirror docs.

agentbreeder secret migrate

agentbreeder secret migrate --from BACKEND --to BACKEND [--prefix PREFIX] [--include KEY] [--exclude KEY] [--dry-run] [--json]
OptionRequiredDescription
--fromYesSource backend (env, aws, gcp, vault)
--toYesTarget backend (aws, gcp, vault)
--prefixNoPrefix for secrets in cloud backend (default: agentbreeder/)
--include, -iNoOnly migrate these keys (repeatable)
--exclude, -eNoSkip these keys (repeatable)
--dry-runNoPreview without writing

Migrates secrets from one backend to another. After migration, update agent.yaml to use secret:// references:

deploy:
  secrets:
    - OPENAI_API_KEY     # resolved from secret://OPENAI_API_KEY at deploy time

Examples:

agentbreeder secret list                                    # List env secrets
agentbreeder secret list --backend aws --json               # List AWS secrets as JSON
agentbreeder secret set OPENAI_API_KEY                      # Prompt for value
agentbreeder secret set OPENAI_API_KEY --value sk-...       # Provide value directly
agentbreeder secret get OPENAI_API_KEY --reveal             # Print actual value
agentbreeder secret delete OPENAI_API_KEY --force           # Delete without confirmation
agentbreeder secret rotate OPENAI_API_KEY                   # Prompt for new value
agentbreeder secret migrate --from env --to aws --dry-run   # Preview migration
agentbreeder secret migrate --from env --to aws             # Migrate to AWS
agentbreeder secret migrate --from env --to gcp --exclude DEBUG --exclude LOG_LEVEL

agentbreeder schedule

Run an agent on a cron schedule (requires pip install 'agentbreeder[schedule]').

agentbreeder schedule [AGENT_DIR] [--cron EXPR] [--once] [--dry-run]
Argument / OptionRequiredDefaultDescription
AGENT_DIRNo.Path to agent directory containing agent.yaml
--cronConditionalStandard cron expression, e.g. "0 8 * * *". Required unless --once.
--onceNofalseRun immediately once and exit (ignores --cron).
--dry-runNofalsePrint next 5 fire times and exit without running.

Examples:

agentbreeder schedule                          # Run cron schedule from current dir
agentbreeder schedule ./support-agent --cron "0 9 * * 1-5"  # Weekdays at 9am
agentbreeder schedule --once                   # Run once immediately
agentbreeder schedule --dry-run --cron "0 * * * *"  # Preview hourly fire times

agentbreeder eval

Run agent evaluations, inspect datasets, and gate deployments on score thresholds.

agentbreeder eval [SUBCOMMAND] [OPTIONS]

agentbreeder eval run

agentbreeder eval run AGENT_NAME --dataset DATASET_ID [OPTIONS]
OptionRequiredDescription
--dataset, -dYesDataset ID to run against
--model, -mNoModel override for the run
--temperature, -TNoTemperature override
--judgeNoLLM judge model (claude-*, gpt-*, gemini-*)
--scorerNoScoring strategy: exact, semantic, or judge
--jsonNoJSON output

agentbreeder eval datasets

List available evaluation datasets.

agentbreeder eval datasets [--team TEAM] [--json]

agentbreeder eval results

Show results for a specific eval run.

agentbreeder eval results RUN_ID [--json]

agentbreeder eval gate

Assert a minimum score — useful in CI pipelines before promoting a deployment.

agentbreeder eval gate RUN_ID [--threshold 0.7] [--metrics METRICS] [--json]
OptionDefaultDescription
--threshold, -t0.7Minimum acceptable score (0.0–1.0). Exits non-zero if below.
--metricsComma-separated list of metrics to check

agentbreeder eval compare

Compare two eval runs and flag regressions.

agentbreeder eval compare RUN_A RUN_B [--regression-threshold 0.05] [--json]

Examples:

agentbreeder eval run support-agent --dataset support-qa-v2 --scorer judge
agentbreeder eval datasets --team customer-success
agentbreeder eval results run-abc123 --json
agentbreeder eval gate run-abc123 --threshold 0.85        # Fail CI if score < 85%
agentbreeder eval compare run-baseline run-candidate

agentbreeder template

Manage agent templates — browse the community catalog and scaffold agents from templates.

agentbreeder template [SUBCOMMAND] [OPTIONS]

agentbreeder template list

agentbreeder template list [--category CATEGORY] [--framework FRAMEWORK] [--status STATUS]

agentbreeder template create

Publish an existing agent as a reusable template.

agentbreeder template create NAME [OPTIONS]

agentbreeder template use

Scaffold a new agent from a template.

agentbreeder template use NAME [OPTIONS]

Examples:

agentbreeder template list                              # All templates
agentbreeder template list --framework langgraph        # Filter by framework
agentbreeder template list --category support           # Filter by category
agentbreeder template use customer-support-starter      # Scaffold from template

agentbreeder compliance

Generate compliance evidence report packs for SOC 2 Type II, HIPAA, GDPR, and ISO 27001.

agentbreeder compliance [SUBCOMMAND] [OPTIONS]

agentbreeder compliance generate

Export a compliance report pack from the audit log and registry.

agentbreeder compliance generate [--standard STANDARD] [--since SINCE] [--format FORMAT] [--output PATH]
OptionDefaultDescription
--standard, -sallStandards to include: soc2, hipaa, gdpr, iso27001
--since90dLookback window (e.g., 90d, 1y)
--formatjsonOutput format: json or pdf
--output, -o./compliance-reportOutput path

Examples:

agentbreeder compliance generate                          # All standards, last 90 days
agentbreeder compliance generate --standard soc2          # SOC 2 only
agentbreeder compliance generate --since 1y --format pdf  # 1-year PDF report

agentbreeder studio

Start AgentBreeder Studio and the API server via Docker Compose.

agentbreeder studio [--follow] [--pull] [--api-port PORT] [--port PORT]
OptionDefaultDescription
--follow, -fOffTail logs after start (Ctrl+C stops the stack)
--pull / --no-pull--no-pullPull latest images before starting
--api-port8000Local port for the API server
--port3001Local port for Studio

agentbreeder up

Start the full AgentBreeder platform stack (API + Studio + Postgres + Redis + migrations).

agentbreeder up [--no-input] [--no-browser] [--env-file PATH] [--build]
OptionDefaultDescription
--no-inputOffSkip interactive prompts, use defaults
--no-browserOffDon't open browser after startup
--env-filePath to an existing .env file
--buildOffBuild images from source (instead of using cached images)

agentbreeder down

Stop the AgentBreeder platform. Works from any directory — stops the quickstart stack if running, or the dev stack if a compose file is found nearby.

agentbreeder down [--clean] [--json]
OptionDefaultDescription
--cleanOffAlso remove volumes (full reset — deletes all database data)
--jsonOffJSON output
agentbreeder down            # stop everything, keep data
agentbreeder down --clean    # stop + wipe all volumes

Global Options

All commands support:

OptionDescription
--jsonOutput as JSON (for CI/scripting)
--helpShow help with examples

Exit Codes

CodeMeaning
0Success
1General error
2Invalid arguments

On this page

TL;DRGlobal UsageCommandsagentbreeder login / logout / whoamiagentbreeder contextagentbreeder doctoragentbreeder quickstartagentbreeder setupagentbreeder seedagentbreeder initagentbreeder deployagentbreeder validateagentbreeder listagentbreeder describeagentbreeder searchagentbreeder scanagentbreeder registryagentbreeder registry prompt pushagentbreeder registry prompt listagentbreeder registry prompt tryagentbreeder registry tool pushagentbreeder registry tool listagentbreeder registry tool runagentbreeder registry agent pushagentbreeder registry agent listagentbreeder registry agent invokeagentbreeder registry memory pushagentbreeder registry memory listagentbreeder registry rag pushagentbreeder registry rag listagentbreeder registry rag ingestagentbreeder registry rag searchagentbreeder logsagentbreeder statusagentbreeder teardownagentbreeder providerSubcommandsagentbreeder modelSubcommandsagentbreeder chatagentbreeder ejectagentbreeder orchestrationagentbreeder orchestration validateagentbreeder orchestration deployagentbreeder orchestration listagentbreeder orchestration statusagentbreeder orchestration chatagentbreeder submitagentbreeder reviewSubcommandsagentbreeder publishagentbreeder secretagentbreeder secret listagentbreeder secret setagentbreeder secret getagentbreeder secret deleteagentbreeder secret rotateagentbreeder secret syncagentbreeder secret migrateagentbreeder scheduleagentbreeder evalagentbreeder eval runagentbreeder eval datasetsagentbreeder eval resultsagentbreeder eval gateagentbreeder eval compareagentbreeder templateagentbreeder template listagentbreeder template createagentbreeder template useagentbreeder complianceagentbreeder compliance generateagentbreeder studioagentbreeder upagentbreeder downGlobal OptionsExit Codes