compareeHub

BUILD A TALKING AI ASSISTANT ON A BREADBOARD AND HOST ITS BRAIN YOURSELF

A breadboard, five cheap parts, and it talks back to you — with the wake word running on the chip itself.

by 78

FULL CAD BOM FIRMWARE DOCS

AIOpen-hardware

difficulty
●●●○○
time
a weekend
license
MIT
repo
repo ACTIVE28,698 stars

WHAT YOU’LL NEED

  • Electronic parts — full list with part numbers in the repo BOM
  • Dev board / microcontroller — runs the project firmware

Partner

Run your scrapers in the cloud — This build needs data collection that runs 24/7. Apify's free plan includes $5 in credits every month — no card needed.Try Apify free
Videos
1

COMPAREE VERDICT

XiaoZhi is an open-source voice assistant that runs on an ESP32 you wire up yourself on a breadboard. The reference build is a Xiao ESP32-S3, an INMP441 MEMS microphone, a MAX98357 I2S amplifier, any small speaker and a 0.9 inch OLED screen — jumper wires, no custom board. You flash a prebuilt binary and the OLED comes alive with an animated face; from then on you say the wake word and talk to it. Wake-word detection runs offline on the chip itself using Espressif's ESP-SR, with customizable wake words, so the device is not streaming audio until you address it. What makes it more than a toy chatbot is MCP: device-side MCP lets the model actually control the hardware it lives in — speaker, LED, servo, GPIO — and cloud-side MCP extends it to smart home control, operating a PC desktop, knowledge search and email. The honest limitation is that the device needs a network connection and a backend to think: out of the box it pairs with a hosted console. The project's answer to that is a separate MIT-licensed backend, xinnan-tech/xiaozhi-esp32-server with 10.2k stars, which lets you run the whole server side on your own machine instead of somebody else's. The single biggest trap is mistaking the on-chip wake word for a fully offline assistant — only the wake-word detection is local; the reasoning still goes to a server. If you want full control, you will spend an afternoon setting up the self-hosted backend. If you are happy with the hosted console and want a weekend build that just works, this is one of the cleanest ESP32 voice projects available.

NOT IN THE REPO

  • —Prebuilt firmware binaries for 171 variants across 138 board directories — no compilation required for the reference build.
  • —Full wiring diagrams and pin assignments in the docs; no custom PCB, jumper wires are enough.
  • —MIT license, no commercial restrictions.
  • —The device needs a network connection and a backend to think — out of the box it pairs with a hosted console.
  • —A separate MIT-licensed self-hosted backend exists (xinnan-tech/xiaozhi-esp32-server, 10.2k stars) if you do not want to use the hosted one.
  • —Wake-word detection runs offline on the chip using Espressif's ESP-SR, with customizable wake words — audio is not streamed until you address it.

Can I build this?

Printnothing required
BuyESP32 board (Xiao ESP32-S3 is the reference), INMP441 microphone, MAX98357 amplifier, small speaker, 0.9 inch OLED, breadboard and jumper wires
ToolsUSB cable for flashing, network connection (Wi-Fi or wired Ethernet or USB RNDIS or 4G modem depending on board), account for the hosted backend or a machine to run the self-hosted one
Skillsbasic electronics and soldering-free wiring, comfort with flashing firmware binaries and following pin diagrams — no coding required for the reference build
Timea weekend — an hour to wire it, an hour to flash and test, half a day more if you set up the self-hosted backend
Cost$ — the reference parts are under $15 total, the backend is free whether hosted or self-run
SafetyNone beyond ordinary electronics care. Low-voltage USB power, no mains, no lithium cells in the reference build.

Build at your own risk. Projects involve tools, electronics and sometimes mains voltage — follow the creator’s safety notes.

A desktop 3D printer that prints in full colorPartner · Kickstarter
A desktop 3D printer that prints in full color

HeyGears G1: 10M+ colors and transparent parts in one print, plus UV printing on flat objects. Figures, parts and labels — no painting.

See how it prints

Videos

XiaoZhi breadboard build - Tech Talkies

No official project video; community walkthroughs exist but none are linked in the payload.

Gallery

Start here

Navigation into the creator’s own docs — we don’t rewrite the guide, we route you to the source.

  1. 1.Pick your board(The reference build is a Xiao ESP32-S3, but 138 board directories are supported — check the firmware matrix in the repo for yours.)
  2. 2.Follow the wiring diagram(Pin assignments are in the docs; the reference build uses jumper wires, no soldering.)
  3. 3.Flash the prebuilt firmware(171 firmware variants are available — download the binary for your board and flash it.)
  4. 4.Configure network and backend(Out of the box it pairs with the hosted console; if you want self-hosted, set up xiaozhi-esp32-server separately.)

Resources

Documentation, files and community threads for this build — we link straight to the original sources and never rehost the creator’s files.

Ultra-thin power bank that fits anywherePartner · Kickstarter
Ultra-thin power bank that fits anywhere

CORE POWER BANK S: 0.2" thin, 2.8 oz, 5000 mAh with 22.5W fast charging. Semi-solid battery, carbon fiber, IPX7 waterproof.

Learn more

KNOWN ISSUES

  • The on-chip wake word is not the same as a fully offline assistant — only wake-word detection runs locally, the reasoning still needs a server.
  • If you do not want to use the hosted backend, you will need to run xiaozhi-esp32-server yourself, which is a separate setup step.
  • The firmware matrix covers 138 boards, but if yours is not listed you will be compiling from source.
  • AEC-capable hardware is required for full-duplex conversation where you can interrupt it mid-sentence — the reference board does not include this.
  • MCP integration is powerful but assumes you are comfortable with the hosted console or willing to configure the self-hosted backend to reach it.
  • The OLED is optional but the animated face is half the charm — skipping it saves $3 and loses the visual feedback.

Is this fully offline?

No. Wake-word detection runs on the chip itself using Espressif's ESP-SR, so it is not streaming audio until you address it, but the reasoning and response generation need a network connection and a backend — either the hosted console or the self-hosted server.

Do I need to compile the firmware?

Not for the reference build or any of the 171 supported variants — prebuilt binaries are available. If your board is not in the list, you will compile from source.

What is the difference between the hosted backend and the self-hosted one?

The hosted console is ready to use out of the box; the self-hosted backend (xinnan-tech/xiaozhi-esp32-server) runs on your own machine and gives you full control, but it is a separate setup step.

Can it control smart home devices?

Yes, through cloud-side MCP — but that assumes you are using a backend that supports MCP integration, either the hosted one or your own.

Community builds

No community builds yet — be the first, we feature the best ones.

Discussion1

FROM THE COMPAREE TEAM

28,700 stars, 171 firmware variants, and the wake word runs on the chip itself — but the reasoning still needs a server. Would you run the self-hosted backend, or is the hosted console good enough for a weekend build?

CompareeTEAM17d ago

Practical notes from our verification: the repo is actively developed with 138 board directories and 171 firmware variants, the reference build uses jumper wires on a breadboard with no custom PCB, and the single biggest decision is whether to use the hosted backend or spend an afternoon setting up the self-hosted one (xinnan-tech/xiaozhi-esp32-server). The on-chip wake word using Espressif's ESP-SR is real and means the device is not streaming audio until you address it, but the reasoning and response generation still need a network connection — this is not a fully offline assistant. The MCP integration is the standout feature: device-side MCP lets the model control the hardware it lives in (screen, LED, servo, GPIO), and cloud-side MCP extends it to smart home control and more. The docs are thorough, the firmware is prebuilt for the supported boards, and the animated OLED face is half the charm. If you want a voice assistant you can wire up in an evening and either trust the hosted backend or run your own, this is one of the cleanest ESP32 projects available.

78

78 built XiaoZhi as an open-source ESP32 voice assistant with on-chip wake-word detection and MCP integration, and released both the firmware and a self-hosted backend under the MIT license. The project has 28.7k stars and 6.5k forks.

GitHub

Star the project on GitHub

DISCLAIMER

  • Comparee is not the author of the projects featured here. All rights to each project belong to its creator — every page links to the original source, and we never host creators’ files.
  • Information is provided without warranty and may become outdated as projects evolve. Prices are indicative bands only — always check the creator’s parts list for current costs.
  • Building and operating any project is at your own responsibility. Protective equipment, safe workshop practice and compliance with local regulations are the builder’s responsibility.
GitHub repo