# AI use cases for Pi Rust (bigfish1913/pi-rust) Repo: [bigfish1913/pi-rust](https://github.com/bigfish1913/pi-rust) | Use cases: 4 | Page: https://dedupedai.com/use-cases/bigfish1913-pi-rust/ rpi is a Rust-native coding-agent runtime with a terminal CLI, built as nine composable crates for LLM providers, tools, sessions, and a plugin ABI. This file holds every use case of this repo. Each one starts with a "## " heading and ends with its URL. - Run a one-shot coding task from the terminal: https://dedupedai.com/use-cases/bigfish1913-pi-rust/run-terminal-coding-agent-one-shot/ - Feed agent events into a script as JSON: https://dedupedai.com/use-cases/bigfish1913-pi-rust/pipe-agent-events-into-a-script/ - Test an embedded agent offline with no API key: https://dedupedai.com/use-cases/bigfish1913-pi-rust/test-agent-tools-offline-in-rust/ - Build a plugin tool that loads into every rpi session: https://dedupedai.com/use-cases/bigfish1913-pi-rust/build-plugin-tool-for-every-session/ ## How to run a one-shot coding task from the terminal with Pi Rust Repo: [bigfish1913/pi-rust](https://github.com/bigfish1913/pi-rust) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Install the rpi CLI and send it a single task from the shell with a model API key. **Who it is for:** A developer who wants a coding agent in the terminal to answer questions about a project or make a small change. **The problem:** Setting up an agent usually means wiring up SDKs or web tools before any work gets done. **The outcome:** A one-line command returns the agent's answer for the task in the terminal. ### Steps 1. Run `cargo install rpi-cli` 2. Run `rpi --version` to confirm the install 3. Run `export ANTHROPIC_API_KEY=...` or `export OPENAI_API_KEY=...` with your own key (inferred, check the README) 4. Run `rpi -p "summarize the README"` from the project directory ### Prompt to paste into your agent ```text Install rpi-cli from the bigfish1913/pi-rust repository using cargo, then run `rpi -p ""` in with the API key set. Expected output: a text answer to the task. Verify that `rpi --version` prints a version before running the task. ``` ### Caveat The README's interactive TUI and sessions need a configured provider, so the first run will not work without a key or a models.json file. Risk, credentials: The agent needs a model API key in your shell environment, so keep the key out of shared scripts and logs. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/bigfish1913-pi-rust/run-terminal-coding-agent-one-shot/ ## How to feed agent events into a script as JSON with Pi Rust Repo: [bigfish1913/pi-rust](https://github.com/bigfish1913/pi-rust) | Difficulty: Intermediate | Time: 30 minutes | Basis: From the README Use the JSON output mode to get a machine-readable event stream from a one-shot run. **Who it is for:** A DevOps or automation engineer who wants to run the agent from a script and parse its output without scraping plain text. **The problem:** Plain terminal output is hard to parse reliably in automation. **The outcome:** A script receives the agent's events as a JSON stream it can process. ### Steps 1. Install the CLI with `cargo install rpi-cli` 2. Set a model key, for example `export OPENAI_API_KEY=...` (inferred, check the README) 3. Run `rpi --mode json -p "list the crates"` and capture stdout 4. Parse the captured stream in your script, one event per line if the output follows the README's description of an event stream ### Prompt to paste into your agent ```text Using the bigfish1913/pi-rust CLI, run `rpi --mode json -p ""` in and save the output to . Expected output: a machine-readable event stream. Verify that the file is non-empty and that each line parses as JSON before using it downstream. ``` ### Caveat The README describes the JSON mode only as a machine-readable event stream, so check the event schema in the user guide before relying on specific field names. Risk, credentials: The script needs a model API key in its environment, so store it in a secret manager rather than in the script. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/bigfish1913-pi-rust/pipe-agent-events-into-a-script/ ## How to test an embedded agent offline with no API key with Pi Rust Repo: [bigfish1913/pi-rust](https://github.com/bigfish1913/pi-rust) | Difficulty: Advanced | Time: 45 minutes | Basis: From the README Run the minimal example against the deterministic faux provider to test agent behavior. **Who it is for:** A Rust developer building an agent into their own app who needs repeatable tests that never touch the network. **The problem:** Agent tests that call a live model are slow, flaky, and cost money to run. **The outcome:** The agent loop runs end to end with fixed output and no credentials, so tests pass the same way on every run. ### Steps 1. Run `git clone https://github.com/bigfish1913/pi-rust.git` 2. Change into the folder with `cd pi-rust` 3. Run `cargo run -p minimal` 4. Read examples/minimal/src/main.rs and reuse its AgentBuilder setup in your own test ### Prompt to paste into your agent ```text Clone bigfish1913/pi-rust, run `cargo run -p minimal`, and then write a Rust test in that builds an agent with the faux provider and asserts the reply. Expected output: the minimal example prints 2 messages after run and the assistant reply. Verify that the test passes with no network access and no API key set. ``` ### Caveat The faux provider covers the paths your test scripts, so live provider behavior still needs its own checks against the real HTTP providers behind the providers feature flag. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/bigfish1913-pi-rust/test-agent-tools-offline-in-rust/ ## How to build a plugin tool that loads into every rpi session with Pi Rust Repo: [bigfish1913/pi-rust](https://github.com/bigfish1913/pi-rust) | Difficulty: Advanced | Time: 90 minutes | Basis: From the README Write a cdylib plugin with rpi-plugin-sdk so the global rpi binary loads your tool. **Who it is for:** A developer who wants a custom tool or provider available in every rpi session without forking the CLI. **The problem:** Adding a tool to a coding agent usually means changing its core code and maintaining a separate fork. **The outcome:** A compiled plugin library that the global rpi loads alongside the built-in tools. ### Steps 1. Run `git clone https://github.com/bigfish1913/pi-rust.git` 2. Open examples/plugin-stub and read how it uses rpi-plugin-sdk 3. Run `cargo build -p plugin-stub` from the repository root to build the cdylib (inferred, check the README) 4. Place the built library where the rpi plugin loader expects it, following docs/agent/project-structure.md ### Prompt to paste into your agent ```text In the bigfish1913/pi-rust repository, copy examples/plugin-stub into , implement one tool in it, and build it as a cdylib. Expected output: a compiled shared library that rpi can load. Verify that `rpi --version` still runs after the plugin is placed in the loader path. ``` ### Caveat The README does not give the plugin install path or build command, so confirm them in examples/plugin-stub and docs/agent/project-structure.md before copying files. Disclaimer: Third-party repos we don't own or vouch for. AI-generated use cases can be wrong: review everything and run at your own risk. Full text: https://dedupedai.com/disclaimer/ URL: https://dedupedai.com/use-cases/bigfish1913-pi-rust/build-plugin-tool-for-every-session/