Crate prompt_store 

prompt-store: A secure, encrypted CLI and library for managing AI prompts.

This crate provides both a command-line interface and a Rust library for securely storing, organizing, and executing AI prompts and multi-step prompt chains.

Library Usage Example

use prompt_store::{PromptStore, RunOutput};
use llm::builder::{LLMBuilder, LLMBackend};
use llm::chain::LLMRegistry;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    // Initialize the store once. This handles loading the encryption key.
    // For non-interactive environments, prefer `PromptStore::with_password("my-secret")`.
    let store = PromptStore::init()?;

    // Setup LLM providers
    let openai_key = std::env::var("OPENAI_API_KEY")?;
    let openai_llm = LLMBuilder::new()
        .backend(LLMBackend::OpenAI)
        .api_key(openai_key)
        .model("gpt-4o-mini")
        .build()?;

    let mut registry = LLMRegistry::new();
    registry.insert("openai", openai_llm);

    // --- Run a single stored prompt ---
    let output = store.prompt("my-prompt-id")
        .vars([("name", "Alice")])
        .backend(registry.get("openai").unwrap())
        .run()
        .await?;

    if let RunOutput::Prompt(text) = output {
        println!("Single prompt output: {}", text);
    }

    // --- Define and run a dynamic chain ---
    let chain_output = store.chain(&registry)
        .step("analysis", "prompt-for-analysis") // Loads a prompt with ID/title "prompt-for-analysis"
            .with_provider("openai")
        .step_raw("synthesis", "Synthesize the following: {{analysis}}") // Uses a raw string as a prompt
            .with_provider("openai")
        .vars([("input_data", "Some data to analyze.")])
        .run()
        .await?;

    if let RunOutput::Chain(map) = chain_output {
        println!("Final chain output: {}", map.get("synthesis").unwrap());
    }

    Ok(())
}

Re-exports

pub use api::PromptStore;

pub use api::RunError;

pub use api::RunOutput;

pub use api::StoreError;

Modules

api

High-level fluent API for running prompts and chains.

cli

Defines the command-line interface structure using clap.

commands

core

ui