Show HN: Hey Spark – open-source voice assistant for your laptop, any LLM brain

ContextuAI released Hey Spark, an open-source always-listening voice assistant for macOS 14+ on Apple Silicon and Windows 11 that runs wake-word detection (openWakeWord), speech-to-text (faster-whisper) and text-to-speech (Piper or Kokoro) locally on the CPU while routing requests to Claude via the Claude Agent SDK, Hermes Agent, or any OpenAI-compatible model including local ones through Ollama. Hey Spark connects to tools such as GitHub, Gmail and Calendar through MCP servers with per-server allowlists, requiring a spoken "yes" before any tool that sends, creates or updates, and blocking destructive tools by default. The release requires about 1 GB of disk for models, and audio leaves the machine only if a user selects a cloud STT backend such as Deepgram or ElevenLabs.

by ContextuAI https://contextuai.com An always-listening voice assistant for your laptop. Say "Hey Spark" , ask a question or give a command, and hear the answer. The wake word, speech recognition and voice all run locally on your CPU. The brain is Claude, Hermes Agent https://github.com/NousResearch/hermes-agent , or any OpenAI-compatible model, including a local one through Ollama. Spark reaches your tools GitHub, Gmail, Calendar, your own systems through MCP https://modelcontextprotocol.io servers, and it asks out loud before doing anything that changes something. "Hey Spark, what's on my calendar tomorrow?" "Set a pasta timer for ten minutes." "Summarize the open pull requests on hey-spark." "Let's start a new project called weather-bot." → "I'll create a folder, a git repo and a private GitHub repo called weather-bot. Should I go ahead?" - Local wake word. A bundled openWakeWord https://github.com/dscripka/openWakeWord model for "Hey Spark". Train your own wake word https://github.com/contextuai/hey-spark/blob/main/docs/wakeword.md for free in Google Colab, or use Picovoice Porcupine. - Local speech. faster-whisper https://github.com/SYSTRAN/faster-whisper speech-to-text and Piper https://github.com/rhasspy/piper or Kokoro https://github.com/thewh1teagle/kokoro-onnx text-to-speech, all CPU-only. Deepgram and ElevenLabs are optional cloud backends. - Your choice of brain. Claude through the Claude Agent SDK https://pypi.org/project/claude-agent-sdk/ , your own Hermes Agent https://github.com/NousResearch/hermes-agent , or any OpenAI-compatible API: OpenAI, Azure OpenAI, OpenRouter, Groq, Together, Ollama, LM Studio or vLLM. - Connectors over MCP , with per-server allowlists. Read-only tools run freely; anything that sends, creates or updates needs a spoken "yes" , and destructive tools are blocked by default. - Conversation that feels natural. Spark streams its answer sentence by sentence, stays awake for follow-ups without the wake word, ignores speech that isn't meant for it, and can be interrupted by saying the wake word barge-in . - A personality of your choice. Plain and brief by default, or a composed British butler with a dry wit who calls you "sir", answers the wake word with "Yes, sir?" and speaks in a British voice. See Personality and voice https://github.com/contextuai/hey-spark/blob/main/docs/persona.md . - Memory and shortcuts. It keeps editable Markdown notes about each connected system, and you can map a short phrase "good morning" to a detailed prompt. - Desktop app for macOS and Windows: a single Start/Stop button, a light that shows what Spark is doing, and windows for Connectors, Shortcuts, the Brain, and Personality & Voice. - Everything is swappable from config.yaml : wake word, VAD, STT, TTS and brain backends, with no code changes. php flowchart LR mic Mic 16 kHz -- wake Wake word<br/ openWakeWord wake -- vad Record until silence<br/ Silero VAD vad -- stt Speech to text<br/ faster-whisper stt -- brain Brain<br/ Claude or OpenAI-compatible brain <-- mcp MCP connectors<br/ + spoken confirmation brain -- split Sentence splitter split -- tts Text to speech<br/ Piper / Kokoro tts -- spk Speaker Nothing leaves your machine before the wake word. After it, only the text of your request goes to the brain you chose. Audio leaves the machine only if you pick a cloud STT backend. Requirements: macOS 14 or later on Apple Silicon, or Windows 11 see Platform support platform-support , a microphone, about 1 GB of disk for models, and a brain: Claude Code https://code.claude.com/docs/en/setup signed in to your Claude account or with an Anthropic API key https://console.anthropic.com/ , an OpenAI-compatible key, or Ollama https://ollama.com running locally. A USB speakerphone with echo cancellation works best. Get the latest build from Releases https://github.com/contextuai/hey-spark/releases . Each file has a .sha256 checksum next to it. macOS Apple Silicon, macOS 14+ 1. Download HeySpark-<version -macos-arm64.dmg , open it and drag HeySpark into Applications . 2. Open it from Applications. If macOS says it can't verify the app, open System Settings → Privacy & Security , scroll down and click Open Anyway . Or run xattr -dr com.apple.quarantine /Applications/HeySpark.app once in Terminal. 3. Allow microphone access when macOS asks. Settings, logs, models and your .env live in ~/Library/Application Support/HeySpark . Windows Windows 11 1. Download and run HeySpark-Setup-<version .exe . It installs per-user and needs no admin rights. The installer isn't code-signed yet, so Windows SmartScreen may warn you: choose More info → Run anyway . 2. Open Hey Spark from the Start menu. Settings, logs, models and your .env live in %LOCALAPPDATA%\HeySpark . Then, on either platform: 1. Click the brain button, pick a provider and enter your key or pick Ollama . 2. Press Start . The first start downloads the speech models about 200 MB . Then say "Hey Spark" . 1. Install uv https://docs.astral.sh/uv/ ; it manages Python 3.11 for you git clone https://github.com/contextuai/hey-spark.git cd hey-spark uv sync --all-extras plain uv sync skips the optional backends 2. Download the speech models into ./.models about 200 MB uv run spark models download 3. Add your brain's key cp .env.example .env Windows: copy then set your key; see "Choose a brain" 4. Check everything, then talk to it uv run spark doctor every required check should show ✔ uv run spark say "Hey Spark, what's the capital of Japan?" No microphone handy? uv run spark --text lets you type questions and read answers, and uv run spark --brain mock --text runs fully offline with no key at all. | Brain | Set up | Good for | |---|---|---| | Claude default | Claude Code https://code.claude.com/docs/en/setup installed and signed in or ANTHROPIC API KEY in .env | The best tool use and connector work. Supports the researcher sub-agent and fast/deep model routing | | Hermes Agent | Hermes Agent https://github.com/NousResearch/hermes-agent with its API server on, and its API SERVER KEY | Giving your Hermes agent a voice: it answers with its own model, tools, memory and skills | | OpenAI-compatible cloud OpenAI, Azure, OpenRouter, Groq, Together | The provider's key in .env , and brain.backend: openai compat | Using a model or account you already have | | Local Ollama, LM Studio, vLLM | Run the server, pull a tool-capable model e.g. qwen2.5:7b | Fully local : with faster-whisper and Piper, nothing leaves your machine | The easiest way to switch is the desktop app's Brain window, which also tests the connection and checks that the model can call tools. See docs/openai brain.md https://github.com/contextuai/hey-spark/blob/main/docs/openai brain.md for the provider presets and the config.yaml settings. With Claude, brain.mode picks the models: auto default sends quick requests to brain.fast model and bigger ones to brain.deep model ; deep answers everything with brain.deep model and hands tool-heavy lookups to a researcher on a faster model the best answers, and the highest API cost . You can also switch by voice "use the deep model", "back to auto" . The default config.yaml uses the bundled model assets/wakewords/hey spark.onnx . Nothing to set up. If it misses you or wakes on its own, tune wakeword.options.openwakeword.threshold . Want your own wake word? tools/wakeword trainer.ipynb https://github.com/contextuai/hey-spark/blob/main/tools/wakeword trainer.ipynb trains one for any phrase on a free Colab GPU in about an hour, with no recordings needed. docs/wakeword.md https://github.com/contextuai/hey-spark/blob/main/docs/wakeword.md walks through it step by step, plus the Porcupine alternative. - Follow-ups: after a reply, Spark stays awake for 3 minutes attention.awake s , so you can keep talking without the wake word. It answers only speech that's meant for it, then plays a soft chime when it falls asleep. - Commands: "new topic" fresh conversation , "go to sleep", "never mind", and "mute yourself" it then stops listening until you click the mic button in the window . Film-style ones work too: "stand down", "belay that", "clean slate protocol", "full power", "go dark". - Timers and time: "set a pasta timer for 10 minutes", "what timers do I have", "what time is it". - New projects: "let's start a new project called weather-bot". After your yes , Spark creates projects.root dir/weather-bot with git and a first commit, a private GitHub repo and a tracking issue through your gh login it never pushes , and can open Claude Code in the folder. - Shortcuts: map a phrase like "good morning" to a detailed prompt in the Shortcuts window or shortcuts.json . Fifteen come built in: "rise and shine" for the morning briefing, "what's next?", "sitrep", "check the post", "run a diagnostic", "house party protocol" and more. See docs/commands.md https://github.com/contextuai/hey-spark/blob/main/docs/commands.md . - Barge-in: with app.barge in: true , saying the wake word while Spark is talking stops it. Turn this on only with an echo-cancelling speakerphone. - Earcons: a chime on wake, a tick while thinking, a low tone on errors. - Personality and voice: pick the butler personality and a British voice in the Personality & Voice window or app.persona: butler in config.yaml . See docs/persona.md https://github.com/contextuai/hey-spark/blob/main/docs/persona.md . Connectors are MCP servers defined in mcp servers.json with ${VAR} placeholders only; the values come from .env . Add them by hand or in the desktop app's Connectors window. Each server's tool policy lives in config.yaml under connectors: : - allowed tools are read-only and pre-approved. - side effecting tools send, create, update need a spoken "yes" every time. Spark says one line about what it will do and waits. Silence, an unclear answer or "no" means it doesn't happen. - destructive tools are blocked unless you set allow destructive: true for that server. - Any tool on none of these lists is denied. 1. Create a fine-grained token https://github.com/settings/personal-access-tokens/new for the repos Spark may see: Issues: Read and write , Pull requests: Read and Contents: Read . 2. In the Connectors window, open github under From config files and click Edit config . Replace ${SPARK GITHUB PAT} with your token and press Save and test . Spark moves the token to .env . { "github": { "type": "http", "url": "https://api.githubcopilot.com/mcp/", "headers": { "Authorization": "Bearer YOUR TOKEN" } } } 3. Switch it on. Spark uses GitHub's own labels for the tools: reading runs freely, creating an issue asks you first, and deleting is off. Try "What are the open issues in my-repo?" This is GitHub's official github-mcp-server https://github.com/github/github-mcp-server , hosted by GitHub. GitHub's docs list Copilot access as a prerequisite for the hosted server Copilot has a free plan ; if your account is refused, or you'd rather run it on your own computer, see docs/connectors.md https://github.com/contextuai/hey-spark/blob/main/docs/connectors.md . Paste the token itself: GitHub's VS Code example uses ${input:...} , a VS Code prompt that Spark can't show. Choose Add connectors in the Connectors window and paste any MCP server's JSON the mcpServers block from its README or from Claude Desktop , or import Claude Desktop's servers in one click. docs/connectors.md https://github.com/contextuai/hey-spark/blob/main/docs/connectors.md covers servers set up by hand in mcp servers.json and config.yaml , including Gmail, Google Calendar and Outlook through Composio. The claude.ai Gmail and Calendar connectors are off by default: they work only through a claude.ai login in your own Claude CLI, so they're for personal setups. - Before the wake word, audio never leaves the wake-word detector, and no audio is stored unless you run with --debug-audio . - After the wake word, your words are transcribed locally unless you choose Deepgram and the text goes to the brain you chose. - Secrets live only in .env or the environment. A redaction filter masks them, and common token shapes, in every log line. - Spark keeps local logs you can read and delete: transcripts logs/transcripts.log , turn off with logging.transcripts: false and an activity history history/ , see activity: in config.yaml . They never leave your machine. Everything is in config.yaml https://github.com/contextuai/hey-spark/blob/main/config.yaml , and every key has a comment. Override any key per run with an environment variable SPARK <SECTION <KEY or a CLI flag: wakeword: { backend: openwakeword } openwakeword | porcupine stt: { backend: faster whisper } faster whisper | deepgram tts: { backend: kokoro } kokoro | piper | elevenlabs | print brain: { backend: claude agent } claude agent | openai compat | mock $env:SPARK STT BACKEND = "deepgram"; uv run spark uv run spark --stt deepgram --tts kokoro All commands uv run spark voice mode uv run spark --text type questions, print answers no audio devices uv run spark --text --speak type questions, hear answers uv run spark --no-tts voice in, printed answers out uv run spark --debug-audio also save the last utterance to debug audio/ uv run spark -v debug logging shows tool calls by name uv run spark desktop the desktop window uv run spark doctor environment checks, each with a fix uv run spark models download whisper + the configured voice Kokoro George by default uv run spark bench stt|tts|brain latency benchmarks median / p90 uv run spark bench latency live: wake word to first audio, end to end uv run spark bench false-wake --minutes 30 live: count false wakes on background audio Per-turn latency breakdowns are appended to logs/latency.jsonl . | Platform | Status | |---|---| | macOS 14+, Apple Silicon | Desktop app .dmg and source. Every build is self-tested in CI. New in 0.1.0, so please report anything odd. | | Windows 11 | Desktop app installer and source. The platform Hey Spark was developed on WASAPI audio . | | Intel Macs, Linux | From source only, untested. The code is cross-platform Python; reports and pull requests are welcome. | Everything runs on the CPU; no GPU is needed. macOS on a Mac with the Xcode command-line tools : bash scripts/build macos.sh This freezes the app into build/dist/HeySpark.app , signs it ad hoc, runs its self-test and writes build/installer/HeySpark-<version -macos-arm64.dmg . Set MACOS SIGN IDENTITY plus APPLE ID , APPLE TEAM ID and APPLE APP PASSWORD to sign with a Developer ID and notarize. Windows: powershell -ExecutionPolicy Bypass -File scripts\build windows.ps1 This freezes the app into build\dist\HeySpark\HeySpark.exe . With Inno Setup 6 https://jrsoftware.org/isinfo.php installed winget install JRSoftware.InnoSetup , it also writes build\installer\HeySpark-Setup-<version .exe . Releases are built by GitHub Actions build-apps.yml https://github.com/contextuai/hey-spark/blob/main/.github/workflows/build-apps.yml , which also builds both apps for pull requests that touch packaging or the desktop app. uv run pytest -q offline suite: fakes and synthesized audio, about a minute uv run pytest -q -m live hits real Claude / cloud APIs needs keys uv run ruff check . && uv run ruff format --check . The architecture in one line: every component is an abstract class in spark/interfaces.py , and spark/registry.py maps the backend names in config.yaml to implementations. spark/app.py is the state machine. CONTRIBUTING.md https://github.com/contextuai/hey-spark/blob/main/CONTRIBUTING.md explains how to add a backend, run the tests and send a pull request, and docs/design/SPEC.md https://github.com/contextuai/hey-spark/blob/main/docs/design/SPEC.md is the original design spec. Hardware checks that need a real mic are in docs/dev/MANUAL QA.md https://github.com/contextuai/hey-spark/blob/main/docs/dev/MANUAL QA.md . Hey Spark was built with Claude Code https://claude.com/claude-code . The project's agents and skills are in .claude/ https://github.com/contextuai/hey-spark/blob/main/.claude , and the original phase plans and reports are in docs/dev/history https://github.com/contextuai/hey-spark/blob/main/docs/dev/history . Contributions are welcome: see CONTRIBUTING.md https://github.com/contextuai/hey-spark/blob/main/CONTRIBUTING.md and the Code of Conduct https://github.com/contextuai/hey-spark/blob/main/CODE OF CONDUCT.md . Please report vulnerabilities privately as described in SECURITY.md https://github.com/contextuai/hey-spark/blob/main/SECURITY.md . Release notes are in CHANGELOG.md https://github.com/contextuai/hey-spark/blob/main/CHANGELOG.md . The code is MIT https://github.com/contextuai/hey-spark/blob/main/LICENSE © 2026 Nagendra Rao Jogi. The bundled wake-word model assets/wakewords/hey spark.onnx is licensed separately under CC BY-NC-SA 4.0 non-commercial , because its training data has mixed licenses. See assets/wakewords/LICENSE https://github.com/contextuai/hey-spark/blob/main/assets/wakewords/LICENSE . For commercial use, train your own model on data you're licensed to use, or use Porcupine. Licenses for the dependencies and the downloaded speech models are listed in THIRD PARTY NOTICES.md https://github.com/contextuai/hey-spark/blob/main/THIRD PARTY NOTICES.md . Hey Spark is an open-source project from ContextuAI https://contextuai.com , makers of ContextuAI Solo https://contextuai.com/solo .