Skip to content
Needle 3

Free alternatives to Needle 3

200 free apps like Needle 3, closest match first.

Top picks

  1. P

    physical-ai-neural-reconstruction

    Get

    physical-ai-neural-reconstruction is a router agent skill from NVIDIA's skills collection for NuRec and NRE, NVIDIA's neural reconstruction tools. It does not hold the training or rendering commands itself. Instead it tells the agent which

    Free

  2. O

    OfflineLLM

    Launch

    Private on-device AI chat for Android — runs any GGUF model locally via llama.cpp with ARM-optimised SIMD. Zero network permissions, encrypted settings, biometric lock, tamper detection. + GPU Acceleration

    FreeOpen source

  3. A

    apple-on-device-ai

    Get

    Build private, on-device AI features on iPhone, iPad, and Mac with Foundation Models, Core ML, MLX Swift, or llama.cpp. Use when choosing an Apple-local model

    Free

  4. X

    Xybrid

    Launch

    Run ML models on-device or in the cloud with intelligent routing. Zero infrastructure. Privacy-first, cost-optimized. Any model. Any platform.

    FreeOpen source

  5. J

    jetson-llm-benchmark

    Get

    Jetson LLM Benchmark is a skill for measuring how well a large language or vision-language model performs when served on an NVIDIA Jetson. It runs reproducible benchmarks across vLLM, llama.cpp and Ollama, and each runtime is driven by its

    Free

  6. J

    jetson-inference-mem-tune

    Get

    Jetson Inference Memory Tuning is a skill that recommends an inference runtime and the memory flags to pass to it, based on the Jetson variant and the user's model workload. It supports vLLM, SGLang, llama.cpp and TensorRT Edge-LLM on any N

    Free

More like Needle 3

7
N

nvidia-nemoclaw

NVIDIA NemoClaw plugin for secure sandboxed installation

Get
8
T

TheWhisper

Optimized Whisper models for streaming and on-device use

Launch
9
F

foundation-models-on-device

Apple FoundationModels framework for on-device LLM

Get
10
L

Llamafarm

Deploy any AI model, agent, database, RAG

Launch
11
V

Vui

Vui Nano — a small, context-aware text-to-speech model trained on real

Launch
12
C

Comfyui Triton And Sageattention Installer

Cross-platform installer for Triton and SageAttention on ComfyUI

Launch
13
T

ToolNeuron

Encrypted & Privacy First, On Android Device AI App

Launch
14
P

picoclaw-ai-assistant

Ultra-lightweight AI assistant in Go that runs on $10 hardware

Get
15
W

whatcani.run

Find the best models and how to run them locally

Launch
16
H

huggingface-best

About finding the best, top, or recommended model for a task

Get
17
P

Picollm

On-device LLM Inference Powered by X-Bit Quantization

Launch
18
N

Nativ

Local AI, native to your Mac

Launch
19
N

nemo-automodel-model-onboarding

Guide for onboarding new model architectures into NeMo AutoModel

Get
20
P

physical-ai-infrastructure-setup-and-resilient-scaling

Set up, scale, validate, or harden NVIDIA physical AI infrastructure

Get
21
R

Reins

Local AI chat for iPhone, iPad and Mac

Launch
22
J

jetson-llm-serve

Stand up vLLM or SGLang serving on Jetson

Get
23
P

PhoneClaw

PhoneClaw turns phones into local AI agent runtimes

Launch
24
T

tao-run-inference-service

Start, query, and stop a network-specific TAO inference microservice

Get
25
T

Texts To Transformer

Train a tiny Transformer from scratch on your iMessage history

Launch
26
S

Sherpa Onnx

Speech-to-text, text-to-speech, speaker diarization, speech enhancement

Launch
27
M

Minion

A tiny single-file coding agent for self-hosted models

Launch
28
J

jetson-quick-start

Entry skill for Jetson / IGX BSP customization

Get
29
Q

Qwen3 TTS Apple Silicon

Run Qwen3-TTS text-to-speech locally on Mac (M1/M2/M3/M4)

Launch
30
N

nemotron-speech

Routes NVIDIA Nemotron Speech (Formerly Riva) NIM tasks

Get
31
N

nemo-automodel-launcher-config

Configure NeMo AutoModel job launches for interactive runs, Slurm clusters

Get
32
H

huawei-cloud-openviking-embedding-switch

Switch OpenViking's embedding model to a local llama-server

Get
33
L

LocalAI

Unlock the power of AI with LocalAI, the ultimate AI model management tool

Launch
34
V

vss-deploy-detection-tracking-3d

Deploy and operate the RTVI-CV-3D microservice as MV3DT

Get
35
A

aiq-deploy

Asked to install, deploy, run, validate, troubleshoot

Get
36
T

tao-finetune-huggingface-model

Fine-tune any HuggingFace CV / VLM / LLM model on local NVIDIA GPUs inside an

Get
37
N

Nari Qwen3 TTS

Ultrafast Qwen3-TTS: sub-50 ms time-to-first-audio at 10 requests per second

Launch
38
N

NadirClaw

Source-available LLM router and AI cost optimizer

Launch
39
L

llmfit-hardware-model-matcher

Terminal tool that detects your hardware and recommends

Get
40
D

Devo

Model-neutral agent desktop/runtime for private, enterprise

Launch
41
J

jetson-memory-audit

Measure Jetson DRAM/NvMap usage and verify before/after memory reclamation

Get
42
N

nanochat-llm-training

Train your own GPT-2 level LLM for under $100

Get
43
E

ETOS-LLM-Studio

A native LLM client for iOS & Apple Watch

Launch
44
S

SourceWeft

Open-source, self-hostable alternative to ChatGPT and NotebookLM

Launch
45
U

Uncensored-Local-AI-Multiplatform

A cross-platform Flutter app that runs uncensored, 100% offline open-source…

Launch
46
L

Labml

Monitor deep learning model training and hardware usage from your mobile phone

Launch
47
N

nemotron-customize

Plan, configure

Get
48
N

nemo-automodel-recipe-development

Create and modify NeMo AutoModel training and evaluation recipes

Get
49
T

tao-run-on-brev

Run a TAO training/evaluation/inference container on an NVIDIA Brev GPU instance

Get
50
N

nemo-mbridge-perf-moe-dispatcher-selection

Choose the right MoE token dispatcher

Get

About Needle 3

N

Needle 3

8-29 MB foundation model for tiny devices

Needle 3 is a foundation model from Cactus Compute, built for mobile devices, wearables, robots, smart home devices, cars and microcontrollers. The whole model is a single binary of 8 to 29 MB, based on the company's Simple Attention Network. It is built for three jobs. Tool calls pick the right functions from the ones your app exposes and fill in every argument from what the user said. Structured extraction takes messy text and returns typed fields for a shape you declare, such as an invoice, a booking, a notification or a form. Text embeddings return a vector for a sentence, so an app can search, match and route locally. The company says Needle 3 beats models ten times its size on mobile tool calls and matches models two to three times bigger on extraction, and that it gives up general chat ability to do so. The page lists a browser sandbox where you can try it, and a Python install with pip install cactus-needle.

Category
Productivity
Website
cactuscompute.com