# Taichi

Productive, portable, and performant GPU programming in Python.

Repository: https://github.com/taichi-dev/taichi
Canonical: https://ross.abutalabs.com/products/taichi
Homepage: https://taichi-lang.org
Language: C++
License: Apache-2.0
License Family: permissive
Topics: taichi, differentiable-programming, gpu-programming, sparse-computation, gpu, computer-graphics
Last push: 2026-07-06T19:33:09+00:00
Link (homepage): https://taichi-lang.org

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 91, release rhythm 9, longevity 100
- inputs: {"age_days": 3569, "days_push": 58, "days_rel": 398, "gap_med": 220, "n_releases_24m": 2}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 28345, forks 2384 (observed 2026-08-28T04:11:48.276323+00:00)

## What it is
Taichi Lang is an open-source domain-specific language embedded in Python for high-performance parallel numerical computation, using a JIT compiler to offload compute-intensive code to GPUs or CPUs. It supports differentiable programming and spatially sparse data structures, with applications in physical simulation, rendering, computer vision, and AI.

## Use cases
- accelerate compute-intensive python code on gpu
- write differentiable physical simulations like mpm and fluids
- build gpu path tracing and voxel renderers
- implement lidar slam and computer vision pipelines
- run lattice boltzmann airflow simulations with sparse data structures
- prototype high-performance numerical kernels without writing cuda

## When to choose
- you want near-CUDA performance from python-like syntax
- you need differentiable simulation or gradient-based optimization
- you need portable gpu/cpu deployment of numerical code
- you work on graphics, robotics, or simulation research

## When to avoid
- you need general-purpose python acceleration beyond numerical kernels
- you require a mature general-purpose programming language ecosystem
- your workload is simple data processing better served by numpy or pandas
- you need long-term commercial support guarantees

## Facets
- artifact type: library
- maturity: active
- function: compiler, gpu-computing, simulation, graphics, machine-learning, math
- domain: simulation, gpu-computing, machine-learning, computer-vision, programming-languages
- platform: python, cpp, cross-platform, windows
- tags: dsl, jit-compiler, differentiable-programming, sparse-computation, physical-simulation, numerical-computation, embedded-language, computer-graphics, gpu, linux, macos

## Member repositories
- taichi-dev/taichi (main) score 64
- taichi-dev/difftaichi (examples) score 32

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:48.276323+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T16:54:24.581475+00:00, confidence not recorded.
  - readme: https://github.com/taichi-dev/taichi (fetched 2026-08-28T04:11:48.276323+00:00, sha e07977306fb1)
  - homepage: https://taichi-lang.org (fetched 2026-08-29T07:51:15.124692+00:00, sha 1cfee0104bfa)
- Data as of 2026-08-30T08:39:29.467469+00:00.