autonomy deploy

Build or run AI agent deployments.

This command group consists of a number of functionalities for building AI agent deployments, run locally stored AI agent deployments, and run AI agent deployments defined in the on-chain protocol. See the appropriate subcommands for more information.

Options

--env-file FILE

File containing environment variable mappings

Examples

If you have an .env file in the working directory, autonomy deploy will load the .env file automatically. If the file is not present in the working directory you can provide the path to the file using --env-file flag

autonomy deploy --env-file <path_to_dotenv> COMMAND [ARGS]

For loading the environment variables you can use a json file as well. While using a json file you can either use json serialized strings like

env.json

{
    "ALL_PARTICIPANTS": "[\"0x0000000000000000000000000000000000000000\"]"
}

Or the json objects

env.json

{
    "ALL_PARTICIPANTS": ["0x0000000000000000000000000000000000000000"]
}

The framework will handle both of these cases automatically.

autonomy deploy --env-file <path_to_json> COMMAND [ARGS]

autonomy deploy build

Build an AI agent deployment.

This command must be executed within a service folder. That is, a folder containing the AI agent configuration file (service.yaml). The deployment will be created in the subfolder ./abci_build_*.

Usage

autonomy deploy build [OPTIONS] [KEYS_FILE]

Options

--o PATH

Path to output directory.

--n INTEGER

Number of agent instances.

--docker

Use docker as a backend.

--kubernetes

Use kubernetes as a backend.

--dev

Create development environment.

--log-level [INFO|DEBUG|WARNING|ERROR|CRITICAL]

Logging level for runtime.

--packages-dir PATH

Path to packages directory (Use with dev mode).

--open-aea-dir PATH

Path to open-aea repo (Use with dev mode).

--aev

Apply environment variable when loading AI agent config.

--use-hardhat

Include a hardhat node in the deployment setup.

--use-acn

Include an ACN node in the deployment setup.

-ltm, --local-tm-setup

Use local tendermint chain setup.

--image-version TEXT

Define runtime image version.

--agent-cpu-request FLOAT

Set agent instance CPU usage request.

--agent-memory-request INTEGER

Set agent instance memory usage request.

--agent-cpu-limit FLOAT

Set agent instance CPU usage limit.

--agent-memory-limit INTEGER

Set agent instance memory usage limit.

--number-of-services INTEGER Number of AI agents (i.e., separate deployment folders) to build. Default is 1.

--remote

To use a remote registry.

--local

To use a local registry.

-p

Ask for password interactively.

--help

Show the help message and exit.

Examples

autonomy deploy build keys.json -ltm

Builds an AI agent deployment using the keys stored in the file keys.json and applying environment variables to the AI agent configuration file. The deployment will be generated by default for as many agents as keys are stored in keys.json. By default, the command searches for the file keys.json, if no file name is provided.

You can build multiple AI agent deployments at once using the --number-of-services parameter. Each AI agent will be created in its own abci_build_* folder:

autonomy deploy build --number-of-services 2 keys.json

Note: Use --n to set the number of agent instances per AI agent, and use --number-of-services to build multiple separate AI agents at once.

If you are manually defining keys.json make sure to include ledger property for a key object. This will be used when writing key files to the deployment environment. If you don't specify the property ethereum will be used as default ledger identifier, so if you're running an AI agent against a different ledger (eg. solana) you'll have to specify this property manually like this

[
    {
        "address": "Solanaaddress",
        "private_key": "Solanaprivatekey",
        "ledger": "solana"
    }
]

If the keys are generated using autonomy generate-key command the ledger property will be included automatically.

Private key security

When building deployments, you can use password protected privates keys (eg. generated using autonomy generate-key LEDGER --password PASSWORD) to avoid exposing them.

Docker Compose

In a docker-compose deployment you can just export OPEN_AUTONOMY_PRIVATE_KEY_PASSWORD environment variable before running the deployment like

export OPEN_AUTONOMY_PRIVATE_KEY_PASSWORD=PASSWORD
autonomy deploy run --build-dir PATH_TO_BUILD_DIR

or

export OPEN_AUTONOMY_PRIVATE_KEY_PASSWORD=PASSWORD
docker-compose -f PATH_TO_BUILD_DIR/docker-compose.yaml up

Kubernetes

In a kubernetes deployment, you'll have to export the OPEN_AUTONOMY_PRIVATE_KEY_PASSWORD variable when building the deployment, not when running.

export OPEN_AUTONOMY_PRIVATE_KEY_PASSWORD=PASSWORD
autonomy deploy build (...)

Customize resources

You can customize the resource values for the agent instance nodes using --agent-cpu-limit, --agent-memory-limit, --agent-cpu-request and --agent-memory-request flags or using AUTONOMY_AGENT_MEMORY_LIMIT, AUTONOMY_AGENT_CPU_LIMIT, AUTONOMY_AGENT_MEMORY_REQUEST and AUTONOMY_AGENT_CPU_REQUEST environment variables.

autonomy deploy build --agent-cpu-limit 1.0 --agent-memory-limit 2048 --agent-memory-request 1024

will produce agent instance nodes with following resource constrains

  xyz_abci_0:
    mem_reservation: 1024M
    mem_limit: 2048M
    cpus: 1.0

on docker-compose deployment and

        resources:
          limits:
            cpu: '1.0'
            memory: 2048Mi
          requests:
            cpu: '0.5'
            memory: 1024Mi

on kubernetes deployment.

autonomy deploy run

Run an AI agent deployment locally stored.

This command is a wrapper around docker-compose up to run the AI agent deployment. To execute this command you need to be located within the deployment environment subfolder (./abci_build_*), or specify it with the option --build-dir.

Usage

autonomy deploy run [OPTIONS]

Options

--build-dir PATH

Path where the deployment is located.

--no-recreate

If containers already exist, don not recreate them.

--remove-orphans

Remove containers for AI agents not defined in the Compose file.

--detach

Run AI agent in the background.

--help

Show the help message and exit.

Examples

autonomy deploy run --build-dir ./abci_build_hAsH

Runs the AI agent deployment stored locally in the directory ./abci_build_hAsH.

To provide password for the private keys

export OPEN_AUTONOMY_PRIVATE_KEY_PASSWORD=PASSWORD
autonomy deploy run --build-dir ./abci_build_hAsH

autonomy deploy from-token

Deprecation Warning

This command will be deprecated in the future. Please use autonomy fetch instead.

Run an AI agent deployment minted on-chain protocol.

This command allows to deploy AI agents directly without having the need to explicitly fetch them locally (also known as "one-click deployment"). The command requires the TOKEN_ID which can be checked in the Olas Protocol web app. See the mint an AI agent on-chain guide for more information.

To understand how to use various chain profiles refer to Chain Selection section on the autonomy mint command documentation.

Usage

autonomy deploy from-token [OPTIONS] TOKEN_ID KEYS_FILE

Options

--n INTEGER

Number of agent instances to include in the build.

--skip-image

Skip building images.

--aev

Apply environment variable when loading AI agent config.

--docker

Use docker as a backend.

--agent-cpu-request FLOAT

Set agent instance CPU usage request.

--agent-memory-request INTEGER

Set agent instance memory usage request.

--agent-cpu-limit FLOAT

Set agent instance CPU usage limit.

--agent-memory-limit INTEGER

Set agent instance memory usage limit.

--kubernetes

Use kubernetes as a backend.

--no-deploy

If set to true, the deployment won't run automatically

--use-celo

To use celo chain profile to interact with the contracts

--use-base

To use base chain profile to interact with the contracts

--use-optimism

To use optimism chain profile to interact with the contracts

--use-arbitrum-one

To use arbitrum-one chain profile to interact with the contracts

--use-gnosis

To use gnosis chain profile to interact with the contracts

--use-polygon

To use polygon chain profile to interact with the contracts

--use-ethereum

To use ethereum chain profile to interact with the contracts

--use-custom-chain

To use custom-chain chain profile to interact with the contracts

--use-local

To use local chain profile to interact with the contracts

--help

Show the help message and exit.

Examples

autonomy deploy from-token --use-ethereum 2 keys.json

Runs the AI agent deployment registered with TOKEN_ID=2 on the Ethereum mainnet. The deployment will be run for as many agent instances as keys are defined in the keys.json file.

autonomy deploy stop

Stop a deployment running in background, if you use --detach flag on autonomy deploy run or autonomy deploy from-token the AI agent will run in background. You can stop this AI agent using autonomy deploy stop command.

Usage

autonomy deploy stop --build-dir BUILD_DIR

Options:

--build-dir PATH

Path to the deployment build directory.

--help

Show this message and exit.