5 unstable releases
0.3.2 | Nov 7, 2024 |
---|---|
0.3.1 | Nov 1, 2024 |
0.3.0 | Nov 1, 2024 |
0.2.0 | Oct 31, 2024 |
0.1.0 | Oct 31, 2024 |
#83 in Unix APIs
202 downloads per month
110KB
2.5K
SLoC
Robopages Server
Robopages are YAML based files for describing tools to large language models (LLMs). They simplify the process of defining and using external tools in LLM-powered applications. By leveraging the robopages-cli
function calling server, developers can avoid the tedious task of manually writing JSON declarations for each tool. This approach streamlines tool integration, improves maintainability, and allows for more dynamic and flexible interactions between LLMs and external utilities.
Pages are loaded by default from the ~/.robopages/
directory (or any folder set in the ROBOPAGES_PATH
environment variable), see the https://github.com/dreadnode/robopages
repository for examples.
Install with Cargo
This is the recommended way to install and use the tool:
cargo install robopages
Pull from Docker Hub
docker pull dreadnode/robopages:latest
Build Docker image
To build your own Docker image for the tool, run:
docker build . -t robopages
Optionally, you can create a bash alias like so:
alias robopages='docker run -v /var/run/docker.sock:/var/run/docker.sock -v ~/.robopages:/root/.robopages -p 8000:8000 robopages'
Note about Docker
If you are using robopages
inside a container, make sure to share the docker socket from the host machine with the container:
docker run -it \
# allow the container itself to instrument docker on the host \
-v/var/run/docker.sock:/var/run/docker.sock
# share your robopages
-v$HOME/.robopages:/root/.robopages \
# the rest of the command line
robopages view
Build from source
Alternatively you can build the project from source, in which case you'll need to have Rust and Cargo installed on your system and clone this repository.
To build the project:
cargo build --release
The compiled binary will be available in the target/release
directory. You can run it directly or add it to your system's PATH:
# Run directly
./target/release/robopages
# Or, copy to a directory in your PATH (e.g., /usr/local/bin)
sudo cp target/release/robopages /usr/local/bin/
Usage
This project consists of a CLI for creating, viewing and serving robopages as a REST API.
CLI
Install robopages:
# install https://github.com/dreadnode/robopages to ~/.robopages/
robopages install
# install a custom repository
robopages install --source user/repo
# install from a local archive
robopages install --source /path/to/archive.zip
View installed robopages:
robopages view
Create a robopage with the preferred template:
# create with the basic template, will run the command in the current shell
robopages create --name my_first_page.yml --template basic
# create with the docker-image template, will use a docker image to run the command
robopages create --name my_first_page.yml --template docker-image
# create with the docker-build template, will build a docker image to run the command
robopages create --name my_first_page.yml --template docker-build
Validate one or more files:
# validate all pages in ~/.robopages
robopages validate
# validate a specific page
robopages validate --path my_first_page.yml
# do not attempt to pull or build containers
robopages validate --skip-docker
Start the REST API:
[!IMPORTANT] While strict CORS rules are enforced by default, no authentication layer is provided. It is highly recommended to never bind this API to addresses other than localhost (as per default configuration).
# this will pre build and pull all containers
robopages serve
# this will build or pull containers on demand
robopages serve --lazy
Execute a function manually without user interaction:
robopages run --function nikto_scan --auto
You can also define variables to be used in the function call:
robopages run -F httpx_tech_detect -A --define target=www.example.com
Repeat for multiple variables:
robopages run -F function_name -A -D target=www.example.com -D foo=bar
Using with LLMs
The examples folder contains integration examples for Rigging, OpenAI, Groq, OLLAMA and Nerve.
Dependencies
~29–46MB
~782K SLoC