---

SPDX-FileCopyrightText: 2026 Knitli Inc.

SPDX-FileContributor: Adam Poulemanos adam@knit.li

SPDX-License-Identifier: AGPL-3.0-or-later

name: cocoindex-rust description: Comprehensive toolkit for developing with the CocoIndex Rust API. Use when building high-performance operators, embedding the engine in Rust applications, or extending the core framework. Covers LibContext management, custom native operators, and direct execution control.

CocoIndex Rust API

Overview

CocoIndex's Rust API provides low-level, high-performance access to the core incremental dataflow engine. Unlike the Python API which focuses on flow definition and orchestration, the Rust API allows for:

1. Embedding the Engine - Run CocoIndex within Rust services.
2. Native Operators - Implement zero-overhead sources, functions, and targets.
3. Deep Integration - Direct access to LibContext, database pools, and execution plans.

Resources

Detailed documentation for each part of the system:

• API Surface: Overview of modules and crate structure.
• Setup & Context: How to initialize LibContext, settings, and access flows.
• Sources: Implementing SourceFactory and SourceExecutor for custom data ingestion.
• Functions: Implementing SimpleFunctionFactory for high-performance transformations.
• Types: Understanding Value, KeyPart, and schema systems.

When to Use This Skill

Use when users request:

• "Integrate CocoIndex into a Rust application"
• "Implement a high-performance custom source in Rust"
• "How do I write a native Rust function for CocoIndex?"
• "Access internal flow state or database pools from Rust"
• "What is LibContext?"
• "Explain the Value enum in CocoIndex Rust"

Common Usage Patterns

1. Initialization

See resources/api_setup.md for full details.

Copyuse cocoindex::settings::Settings;
use cocoindex::lib_context::create_lib_context;

let settings = Settings::default();
let lib_ctx = create_lib_context(settings).await?;

2. Implementing a Native Source

See resources/api_source.md.

Implement SourceFactory and SourceExecutor to bridge external data systems (Kafka, Postgres, APIs) into CocoIndex efficiently.

3. Implementing a Native Function

See resources/api_function.md.

Implement SimpleFunctionFactory for computationally intensive tasks where Python overhead is undesirable.