Palette Neat Official Release

Hello everyone,

We are thrilled to announce the official launch of Palette Neat, our new AI agentic-driven application development framework for the SiMa.ai Modalix MLSoC platform! :rocket: We are officially releasing versions 0.1.0 and 0.2.0 today (June 16, 2026).

Palette Neat sits right at the application layer of the SiMa.ai software stack, taking the complexity out of building, testing, and maintaining edge AI applications. Under the hood, it leverages powerful GStreamer-based execution, but exposes a much higher-level, intuitive model for developers.

Whether you prefer a hands-on coding approach or agent-assisted development, Neat has you covered.

:sparkles: What Palette Neat Brings to the Table

  • Intuitive APIs: A streamlined Python and C++ programming model tailored for composing edge AI applications.
  • Simplified Workflows: Features a clean Model, Graph, and Run workflow to effortlessly transition from a compiled model package straight to runtime execution.
  • Reusable Application Patterns: Don’t start from scratch! Neat provides out-of-the-box patterns for common workloads like object detection, instance segmentation, tracking, and depth estimation.
  • Robust Development Tooling: Comes equipped with the Neat Development Environment and Model Compiler tooling specifically for building against SiMa.ai targets.
  • AI Agentic Playbooks: Includes a curated set of coding-agent playbooks (sima-neat/playbooks) to accelerate and automate common Neat development workflows.
  • Advanced Diagnostics: Built-in workflow tools and runtime inspection (via our insight visualizer) for validating pipelines and easily debugging application behavior.

:hammer_and_wrench: Dive Into the Ecosystem

Palette Neat is broken down into a few core repositories to keep your development modular and organized. Check out our open-source tools directly on GitHub:

  • :books: Neat Core: The main Neat Library APIs, tutorials, and documentation source.
  • :laptop: Neat SDK: Your dedicated Development Environment for integrated application dev and cross-compilation.
  • :brain: Model Compiler: Tooling to prep your models for SiMa.ai hardware.
  • :mobile_phone: Apps: Example applications and reference pipeline patterns to get you started quickly.
  • :robot: LLima: Runtime and compile-time tooling specifically designed for LLM and VLM workloads.

:globe_with_meridians: Important Updates to Our Web Domains

As we expand our tools like Palette Neat, we are also reorganizing our web properties to make finding resources easier and to better support our growing ecosystem.

Please update your bookmarks with our new domain structure:

  • :open_book: Current Documentation: developer.sima.ai (This is the new home for all of our active documentation, previously hosted at docs.sima.ai)
  • :speech_balloon: Developer Community: community.sima.ai (This is the new home for our community forums and discussions, previously hosted at developer.sima.ai)
  • :file_cabinet: Legacy Documentation: docs.sima.ai (This domain will now serve as an archive for our older, legacy documentation)

:link: Get Started Today

We designed Neat specifically for teams building production-ready AI applications at the edge. Stop hand-maintaining brittle pipelines and start reusing tested patterns!

:backhand_index_pointing_right: Explore the SiMa.ai Neat GitHub Organization
:backhand_index_pointing_right: Read the Official Developer Documentation

We can’t wait to see what you build with Neat! If you have any questions or feedback, drop a comment below.

β€” The SiMa.ai Team

1 Like