Crate pipelin3r 

Pipeline orchestration for LLM-powered workflows.

pipelin3r provides a high-level builder API for orchestrating LLM agent invocations, shell commands, and data transforms in a pipeline. It wraps the shedul3r-rs-sdk HTTP client with authentication, dry-run capture, and bounded-concurrency batch execution.

Re-exports

pub use agent::AgentBuilder;

pub use agent::AgentResult;

pub use agent::AgentTask;

pub use auth::Auth;

pub use bundle::Bundle;

pub use command::CommandBuilder;

pub use command::CommandResult;

pub use error::PipelineError;

pub use executor::Executor;

pub use model::Model;

pub use model::ModelConfig;

pub use model::Provider;

pub use model::Tool;

pub use pool::run_pool;

pub use template::TemplateFiller;

pub use transform::TransformBuilder;

pub use transform::TransformResult;

Modules

agent

Agent builder for single and batch LLM invocations. Agent builder for single and batch LLM agent invocations.

auth

Per-invocation authentication. Per-invocation authentication for LLM agent invocations.

bundle

Bundle packaging for file transfer. Bundle builder for packaging files to send alongside agent invocations.

command

Shell command execution. Shell command execution wrapper.

error

Typed error enum. Typed error enum for the pipelin3r pipeline orchestrator.

executor

Pipeline executor (SDK client + auth + dry-run). Pipeline executor — wraps the SDK client with authentication and dry-run support.

model

Typed LLM model and provider selection. Typed LLM model and provider selection with TOML-based configuration.

pool

Bounded async concurrency pool. Bounded async concurrency pool.

template

Two-phase template filler. Single-pass template filler with injection protection.

transform

Pure function transforms (stub). Pure function transforms for deterministic data processing.