Skip to content
@sima-neat

SiMa.ai Palette Neat

SiMa.ai Neat

SiMa.ai

SiMa.ai Neat 0.6.0 release highlights

SiMa.ai Neat is an application-development framework for building and running AI applications on the SiMa.ai platform. It provides Python and C++ APIs for working with compiled model artifacts, composing application logic, and running AI pipelines on Modalix DevKits and SiMa.ai SoCs.

Neat sits at the application layer of the SiMa.ai software stack. It uses GStreamer-based execution underneath while giving developers a higher-level model for building, testing, and maintaining edge AI applications.

Start Here

What Neat Provides

  • A Python and C++ programming model for composing AI applications
  • A Model, Graph, and Run workflow for moving from compiled model package to runtime execution
  • Reusable application patterns for object detection, segmentation, tracking, depth estimation, and related edge AI workloads
  • Neat Development Environment and Model Compiler tooling for building against SiMa.ai targets
  • Diagnostics and workflow tools for validating pipelines and inspecting application behavior

Repository Map

Repository Purpose
core Neat Library APIs, tutorials, and documentation source
sdk Neat Development Environment for integrated application development and cross-compilation
model-compiler Model Compiler tooling for preparing models for SiMa.ai hardware
apps Example applications and reference pipeline patterns
sima-cli Developer utilities for setup, assets, and workflow automation
insight Runtime inspection, media routing, and interactive testing for Neat applications
llima Runtime and compile-time tooling for LLM and VLM workloads
playbooks Coding-agent playbooks for common Neat development workflows

Private repositories support internal integrations, build infrastructure, model assets, and fleet-based test execution.

Development Model

Neat is designed for teams building production AI applications at the edge:

  • Build and validate application behavior close to the target runtime
  • Keep model packages, application composition, and runtime execution as separate concerns
  • Reuse tested application patterns instead of hand-maintaining brittle pipelines
  • Support both direct developer workflows and agent-assisted development

For installation, build steps, tutorials, and API references, use the current docs at developer.sima.ai.

Popular repositories Loading

  1. core core Public

    SiMa.ai Palette Neat Library

    C++ 8 2

  2. apps apps Public

    Python 7 3

  3. llima llima Public

    Runtime and compile time tooling for LLM, VLM models

    Python 6 3

  4. sima-cli sima-cli Public

    Utilities helping developers to set things up and download assets such as models

    Python 5 3

  5. insight insight Public

    Performance Visualizer and Interactive Testing Tool for NEAT Apps

    JavaScript 4

  6. sdk sdk Public

    Palette Neat Development Environment for cross compilation with model SDK extension support

    Shell 3

Repositories

Showing 10 of 17 repositories

Top languages

Loading…

Most used topics

Loading…