Changelog

​

StackKits picks your game platform and adds use cases in place; SpeechKit Voice talks to your own model

​

New features

• Choose your game platform. StackKits v0.52.0 installs Calagopus for new Game installations, with Pelican and Pterodactyl as alternatives. The owner key can only read, power, use the console, and edit files. stackkit setup game, stackkit game list, power, and allow work the same on all three, and existing installs keep their platform. See Game.
• Add a use case to a running install. v0.51.0 adds stackkit use-cases add, for example mail, without touching your existing workloads, routes, or storage. It is also available through the StackKits MCP. See stackkit use-cases.
• Move your server to another network. v0.51.7 re-binds a server after it moves to a new network, so its listeners follow the new address. See stackkit network.
• Voice with your own model. SpeechKit v0.76.0 lets voice conversations use your own AI endpoint and model, and keeps the conversation you started in chat, including from the floating panel. See Modes and data paths.
• AssemblyAI and Deepgram conversations. v0.75.5 runs registered-agent Voice sessions on AssemblyAI or Deepgram. Your cloud-processing consent is checked throughout, and AssemblyAI asks separately for recording consent. v0.76.0 adds balanced, lower-latency, and higher-accuracy AssemblyAI profiles.
• New live models. v0.74.0 refreshes the model catalog with GPT-Live on Microsoft Foundry and Gemini Transcribe Live dictation, and removes retired models.

​

Updates

• Home Assistant connector stays on your node. StackKits v0.52.0 issues, reuses, and revokes the Home Assistant agent connector through stackkit setup, and keeps it only in encrypted local storage. See Smart Home.
• Restore from earlier recovery points. v0.52.5 exposes authenticated historical recovery points. See Back up and restore.
• More apps run natively on OpenTofu. v0.51.0 lets Paperless, Mosquitto, Stalwart, Zigbee2MQTT, SearXNG, Roundcube, Euro-Office, Navidrome, Audiobookshelf, ESPHome, Emby, and more apply without the Compose wrapper. Native execution is opt-in per app.
• Secrets stay out of environments and data volumes. v0.51.0 delivers app secrets as files. The Zigbee2MQTT MQTT password and the Pterodactyl app key no longer land in a data volume.
• Managed certificates in Advanced mode. v0.51.7 issues Origin CA certificates for managed kombify.me origins. See stackkit advanced.
• One dictation, one entry. SpeechKit v0.75.1 saves a live dictation as a single Library entry. Hold-to-talk inserts the text once when you release the hotkey.
• Keep the complete dictation. v0.76.4 starts recording while the provider connects and recovers missing stream audio from the recording. Your dictionary words apply in every dictation mode.
• Tabbed settings and a compact voice UI. v0.74.0 organizes settings into tabs. The web voice UI adds a recording indicator, a live transcript, and a microphone check before a session starts.
• Longer Voice sessions. v0.75.3 prefers managed speech providers and allows sessions of up to four hours.
• SDK migrations. v0.74.0 and v0.75.0 include breaking changes for Go SDK and Android keyboard embedders. Read the release notes before you upgrade. See Framework.

​

Bug fixes

• Interrupted restores recover. StackKits v0.52.6 recovers an interrupted restore through the verified original release, and v0.52.5 preserves container file ownership during restore.
• Offsite backups start on a fresh bucket. v0.51.2 initializes the offsite backup repository on an empty bucket.
• Paperless sign-in works. v0.50.6 fixes Paperless-ngx sign-in through PocketID and fixes the Mosquitto broker restart.
• Long Deepgram dictations stay whole. SpeechKit v0.75.0 keeps transcribing past the first pause and transcribes the rest from the local recording if the connection drops.
• Stop means stop. v0.75.0 cuts queued speech on cancel, and Android and web clients wait for the server before sending microphone audio.
• Cleaner voice output. v0.74.1 keeps spaces between streamed words, and v0.74.0 removes the white frame behind transparent overlay windows.

​

StackKits Private AI lab, game servers, and an MCP on every install; SpeechKit meets Microsoft 365

​

New features

• Private AI becomes a local AI lab. StackKits v0.48.0 lets you pick Open WebUI or AnythingLLM as the chat and add web search, document parsing, a Hermes assistant, an OpenHands agent under gVisor, and ComfyUI for images and video. It runs on the CPU or on NVIDIA and AMD GPUs. An optional connector links your homelab model to kombify AI over an outbound connection, so no inbound port opens. See Personal AI.
• Game servers. v0.41.0 adds an optional Game use case with the Pterodactyl Panel. stackkit setup game creates Minecraft Java and Bedrock servers with secure defaults, and v0.43.0 adds Paper, Terraria, and Valheim. Setup checks readiness with a real client join and refuses a server the host memory cannot hold. See Game.
• More apps per use case. v0.47.0 installs Nextcloud, Forgejo, Emby, and Passbolt as alternatives to the defaults. Add-ons install next to a use case: Paperless-ngx, Euro-Office, Navidrome, Audiobookshelf, and ESPHome. See the use-case catalog.
• Mail, your way. v0.47.0 adds Roundcube webmail for a mailbox you already have, plus an optional Stalwart mail server on a dedicated node. See Mail.
• Advanced mode for managed homes. v0.47.0 runs change sets, per-stack drift detection and reconcile, coordinated rollback, and restore drills through Terramate over OpenTofu. Basement installs now default to OpenTofu, and v0.48.0 encrypts its state and plans. See stackkit advanced.
• A StackKits MCP on every installation. v0.46.0 publishes a token-protected /mcp endpoint on each Core, so an AI agent can connect straight to your server. Every public CLI command is now an MCP tool or a listed exception, and commands that would expose a secret stay CLI-only. See MCP connector.
• OS updates from the CLI. v0.47.0 plans and applies OS updates and schedules reboots on a node. See stackkit host.
• Prepared images. v0.50.0 adds stackkit image plan, prepare, and verify to cache the toolchain and core images ahead of time. Each installation still gets its own owner keys and secrets on first boot. See stackkit image.
• Meetings meet Microsoft 365. SpeechKit v0.72.0 imports the Teams transcript or the Microsoft 365 Copilot recap after a Teams meeting, and lets you choose Microsoft Foundry for write-ups with transcript consent. A multi-tenant sign-in lets you switch companies without a new install. See Meeting.
• Bring your own Google voice. v0.73.0 restores Google Cloud speech-to-text, text-to-speech, and Gemini Live as optional providers on your own key. See Modes and data paths.
• OpenAI-compatible audio on the SpeechKit server. v0.73.1 serves POST /v1/audio/transcriptions and POST /v1/audio/speech, so clients such as Open WebUI can use it directly. A local-only profile runs without cloud providers.
• SpeechKit in six languages. v0.73.1 and v0.73.2 localize the Android app, the website, the desktop app, and the tray in English, German, Spanish, Simplified Chinese, Hindi, and Arabic.

​

Updates

• Sign-in for more apps. StackKits v0.48.0 signs Paperless-ngx, Media, and Smart Home in through PocketID and provisions household accounts in Files. v0.38.0 adds stackkit user owner status and activate for the owner passkey. See PocketID.
• Smart Home grows. v0.47.1 installs Mosquitto and Zigbee2MQTT with owner-chosen LAN listeners and device passthrough, and v0.50.0 adopts an existing Home Assistant without replacing its state. See Smart Home.
• Every application gets a route. v0.39.0 gives Media, Smart Home, AI, Dev, and Documents the same HTTPS route and data binding as Photos, Files, and Vault.
• Backups by coverage group. v0.37.0 selects data by config, content, or media. Unset coverage resolves to content, so media is always an explicit choice. See Back up and restore.
• Ubuntu 26.04 by default. v0.50.0 targets Ubuntu 26.04 LTS for new installs and keeps 24.04 compatible. v0.35.0 adds backups on Debian 13. See OS compatibility.
• Compatibility from real runs. v0.33.0 and v0.34.0 grade each kit per environment from automated install-to-restore runs, and v0.41.0 grades Cloud Kit on a public VPS as supported.
• Every command documented. v0.33.0 adds examples to stackkit <command> --help and publishes the full CLI reference. The installer also verifies the downloaded archive against the release checksums.
• Basement defaults to lab.home. v0.47.5 uses lab.home as the default domain, and existing Basement domains migrate with signed recovery.
• Meeting notes as timestamped bullets. SpeechKit v0.71.0 writes notes as timestamped bullets and adds a quick note from the meeting pill.
• Techstack Open Core v0.33.5. The Techstack source and installer are published for v0.26.0 and v0.33.5, and the app adds Spanish, Simplified Chinese, Hindi, and Arabic. See Techstack overview.

​

Bug fixes

• 4 GB servers meet the 4 GB minimum. StackKits v0.45.0 rounds installed memory to the nearest GiB, so a 4 GB VPS is no longer refused by Cloud Kit.
• Upgrades from earlier releases work. v0.39.0 fixes stackkit upgrade and stackkit kit list failing with unexpected EOF. Install a current CLI before you upgrade.
• Config changes reach running services. v0.45.0 recreates a service when its generated config file changes, and the game Panel login no longer loads Google reCAPTCHA.
• Safer recovery. v0.48.0 lets a retried Apply recover an interrupted run and resumes an interrupted Cloud install from any directory. v0.49.0 refuses update planning while an update runs and requires explicit approval for stackkit remove with JSON output.
• Cloud Kit fixes. v0.38.0 makes backups and owner sign-in work on Cloud Kit and names a Let’s Encrypt rate limit with its retry time. v0.48.3 rejects local-only Cloud Kit domains at init.
• Updates relaunch on the first try. SpeechKit v0.72.1 makes the first Install and restart relaunch the app.
• Cleaner German dictation. v0.72.0 stops leftover commas after German filler words, says why a meeting screenshot was not saved, and lets write-ups fall through every model with its own budget.

​

SpeechKit on macOS, Google AI retired, and a StackKits installer that picks the kit

​

New features

• SpeechKit runs on a Mac. v0.70.0 publishes SpeechKit-macOS-arm64.zip beside the Windows assets: an ad-hoc signed Dictation beta for macOS 14+ on Apple Silicon, with microphone capture, text injection, a menu-bar item, the bundled local model, and the cloud speech-to-text providers. One known hole ships with it: the global-hotkey listener installs and receives no key events, so hold-to-talk does not work yet and the menu-bar item is the way in. See Install on macOS.
• Microsoft Foundry, connected three ways. v0.69.0 adds a Connect dialog with a project endpoint and API key, the Azure CLI account already on your machine, or your organisation’s own Microsoft app registration. One connection serves Dictation, speech output, Assist, and the Voice Agent; a Test lists the deployments and voices the project really has.
• A Library with one tab per mode. v0.70.0 gives Transcriptions, Quick Notes, and Voice Sessions their own tabs, loads further back through history a page at a time, expands any row in place, and lets you pin a transcription the way a quick note could already be pinned.
• The StackKits installer picks the kit. v0.32.0 makes install.stackkit.cc detect the host, choose between Basement and Cloud, ask for the owner and the use cases before every install, and offer to install a StackKit once the CLI is in place. A wrong installer URL recovers instead of failing. See CLI walkthrough.
• Household users without hand-built accounts. v0.29.0 provisions household users through PocketID groups, and v0.30.0 adds stackkit user add, list, and remove. See PocketID.

​

Updates

• Google voice adapters are gone. v0.70.2 removes Google speech-to-text, Google text-to-speech, and Gemini Live. A configuration that still names one fails closed instead of silently resolving to something else, and an empty device configuration defaults to AssemblyAI. See Modes and data paths.
• The wake word is hey_kubi. v0.70.2 makes it the default and keeps hey_quby as an alias, so an existing setup keeps working.
• Login gates in front of the homelab routes. v0.31.0 puts TinyAuth ahead of the photos, files, vault, media, smart-home, AI, developer, documents, and Coolify routes, and verification accepts the login redirect as healthy instead of reading it as a failure. See TinyAuth.
• Cloud Kit needs no domain of your own. v0.32.0 defaults it to a kombify.me name when no custom domain is given, and mints an execution-channel SSH key before root login is disabled. See kombify.me.
• Apply waits for the certificate. v0.31.5 waits for issuance before verifying public TLS and internal PKI leaves, and v0.32.0 shows progress while generate and apply run instead of a silent pause.
• These docs follow the current release. SpeechKit download links now resolve to the newest published release rather than one frozen version, the macOS client and Meeting have their own pages, and the release line each page states is reconciled automatically against the public release.

​

Bug fixes

• A macOS first run starts. v0.70.0 fixes a fresh install writing a Windows audio backend into its configuration, refusing to open a microphone with it, and exiting with no message, no log, and no crash report.
• The Voice Agent no longer boots onto a retired provider. v0.70.3 fixes a server whose baked configuration still selected Gemini Live and therefore failed to start after the Google retirement.
• Dictation opens the microphone first. v0.70.1 opens the device before the live-stream handshake instead of after it, and installing an update relaunches the app exactly once.
• Removal cleans up native stacks. v0.32.0 makes stackkit remove uninstall native Compose stacks and find the installer workspace, and v0.31.1 defers application observation to restore activation so a restore no longer checks applications that are not running yet.

​

kombify.me public HTTPS names

​

SpeechKit Meeting Mode and network privacy scopes, StackKits Smart Home and Media

​

New features

• Meeting Mode is a first-class mode. SpeechKit v0.67.0 puts Meeting Mode beside Dictation, Assist, and Voice Agent, with its own settings for capture, review generation, languages, and retention. A compact transparent overlay shows capture health and controls during the call, rolling summaries run without blocking capture, and finishing a meeting produces a durable Meeting Review, a five-sentence Executive Brief, and action items in English, German, Spanish, Chinese, Hindi, and Arabic. GitHub Copilot is an optional review-generation provider behind a separate, revocable transcript-processing grant; the local path needs no cloud credentials. See SpeechKit modes.
• Pick a network boundary and SpeechKit enforces it. v0.67.0 adds a Network Scope setting: all networks, local network only, or this device only. The backend enforces the boundary for speech providers, model downloads, updates, telemetry, server connections, and integrations. Switching to a stricter scope suspends incompatible settings instead of deleting them, and Settings explains what is disabled and why. See SpeechKit overview.
• Screenshots inside a meeting, kept on your machine. v0.68.0 captures a chosen window or the monitor under the cursor with a configurable Ctrl+Alt+S shortcut. Snapshots appear as markers in the transcript timeline, are cited in generated Meeting Reviews, and the compact meeting pill briefly unfolds a preview so you know the right window was captured. No image ever leaves the device.
• Find and export any meeting. The v0.68 patch releases group the meeting list by when a meeting happened, add search across titles, transcripts, notes, and write-ups, render notes as formatted Markdown, and add Copy Markdown and Save .md for the complete write-up.
• Smart Home use case in StackKits. StackKits v0.23.0 adds a Home Assistant workload on Basement, Cloud, and Modern kits. Generate writes the reverse-proxy baseline and a native MCP client fragment, and setup creates the Homelab owner account. See the use-case catalog.
• Media use case with Jellyfin. v0.23.0 adds a Jellyfin workload on the standard and high compute tiers. The media library stays owner-custodied and is deliberately not a StackKits backup source. See Jellyfin.
• Plan a homelab in the browser. v0.23.0 launches the StackKits planner at stackkit.cc/planner: read-only, schema-first tools over the public catalog that hand off an init → validate → resolve → generate → plan sequence to the CLI. Apply remains an explicit step on your machine. See MCP connector.

​

Updates

• Meetings finish on their own. SpeechKit v0.68.0 stops a recording shortly after the calling app releases the microphone, and a notification says when the Meeting Review is ready or why it failed. Hand-started recordings are never auto-stopped.
• Module-local compute profiles in StackKits. v0.23.0 through v0.24.27 move compute, storage, and accelerator profiles onto each module with per-workload alternatives. Required host facts must be observed in inventory before Apply, not inferred from declared hardware. See Choosing a kit.
• The low tier gets real lite variants. v0.23.0 makes a Basement low generate emit standalone Compose variants: core without Coolify and Immich without machine learning. Media is omitted on low until a lite substitute exists, and fails closed at init instead of deploying something undersized.
• Integrations show their maturity. SpeechKit v0.68.0 marks each integration in Settings as Ready, Partial, Planned, or Unavailable, with per-mode tags. What this build cannot run is greyed out instead of failing when enabled.
• A leaner SpeechKit SDK. v0.65.0 removes the provider names deprecated in v0.64.0, so a Deepgram-only app compiles 49 external packages instead of 286. v0.68.0 splits the root package into contracts and implementations with a mechanical import rewrite. v0.66.0 publishes all five Android modules as Apache-2.0 artifacts on JitPack, resolvable anonymously. See SpeechKit framework.
• The overlay settings are a designer. The v0.68 patch releases redraw General settings as compact cards and preview the overlay live: style, position, and the quick buttons on the hover strip update in the preview at once.

​

Bug fixes

• Nothing fails quietly anymore. SpeechKit v0.68.0 reports dropped Voice Agent speech, overlay placement failures, silent screenshot errors, and stalled meeting summaries in the activity log with a reason instead of leaving you guessing.
• Meeting audio stays out of dictation. v0.68.0 runs meeting transcription on its own pipeline. Call audio no longer flashes the dictation overlay, fills the dictation history, or becomes the “last transcription” that the copy and insert shortcuts reuse.
• Long meetings produce write-ups. The v0.68 patch releases size summary batches for the bundled local model, split overlong transcript pieces at sentence boundaries, resume write-ups interrupted by a restart, and retry a temporarily unreachable local model instead of failing the review for good.
• Local dictation starts on non-ASCII install paths. v0.68.0 fixes the bundled whisper-server failing on a fresh install when the model path contains an umlaut. If local speech-to-text still cannot start, the activity log names the cause and says whether dictation is offline or falling back to a named cloud provider.
• Multi-language reviews run one after another. v0.68.0 stops parallel review runs from starving the local model and missing their deadlines together, and progress follows the language actually running instead of a queued job stuck at 0%.
• The dashboard opens at a size that fits. The v0.68 patch releases stop the window from opening taller than a scaled laptop screen, keep the minimum size within a 13-inch display, and preserve a size you choose.
• The Connect button tells the truth. v0.67.0 fixes the hosted connection claiming to be ready without ever obtaining a credential, and the Android app now says when a Companion pairing fails because the two installed apps do not match, instead of showing a general failure.
• StackKits release archives carry their own docs. v0.24.2 includes the public backup and application-readiness documentation in release archives and keeps the guided Basement and Cloud recipes explicit through the compatibility adapter.

​

SpeechKit meeting notes and LAN discovery, StackKits public addresses and compute tiers

​

New features

• Meeting notes that start from your own notes. SpeechKit v0.59.0 adds a meeting notepad: your notes lead, the transcript fills in the rest, and the write-up cites the words it came from. Microphone and system audio stay on separate tracks, so speaker labels come for free and the audio is never kept. SpeechKit notices when a calling app takes the microphone and offers to take notes; nothing is recorded until you accept. The write-up lands in the dashboard as soon as the meeting ends. See SpeechKit modes.
• Find your SpeechKit server on the home network. v0.60.0 lets a homelab server announce itself over mDNS/DNS-SD, so the desktop app, the Android app, and speechkitctl discover find it without typing an address. Discovery is opt-in and off by default, so a hosted instance stays silent, and the announcement only says where to knock: devices still sign in after they connect. See SpeechKit overview.
• Managed public addresses with automatic TLS on Cloud Kit. StackKits v0.21.20 through v0.21.24 give Cloud Kit services managed public kombify.me addresses: a secret-free address plan, routes rendered from your validated subdomain prefix, and certificates issued with ACME HTTP-01 so the Cloudflare gateway can validate origin TLS while service traffic stays HTTPS. See Cloud Kit.
• Compute tiers match the kit to your hardware. StackKits v0.22.0 adds low, standard, and high compute tiers per kit. Apply attests local CPU, RAM, and disk before resolving, and blocks with a clear reason instead of deploying onto an undersized host. See Choosing a kit and Review, plan, and apply.
• Interrupt the assistant from any client. SpeechKit v0.62.0 brings tap-to-stop for a spoken answer to the web and Android clients, not only the desktop app. Transcripts on web and Android carry speaker labels when speaker recognition is on, and each client shows which speech backend is serving the session. See SpeechKit modes.
• Voice-note transcription and summaries on Android. v0.60.0 transcribes voice notes and summarizes WhatsApp and Telegram voice messages from the SpeechKit app.
• Add secrets custody to an existing homelab. StackKits v0.21.11 adds an explicit, idempotent stackkit secrets materialize command so a standalone homelab can establish owner-bound custody after selecting a new workload, without re-running its initial StackSpec. See Day 2 operations.

​

Updates

• Words and Replacements are manageable. SpeechKit v0.62.0 makes custom Words and Replacements searchable, paged, and deletable in Settings. Several spoken forms can rewrite to one term, and aliases apply in Dictation, Assist, and Voice Agent.
• Pick your dictation engine in one control. v0.62.0 switches Deepgram Nova-3 and AssemblyAI Universal 3.5 in one Dictation setting, and AssemblyAI’s LLM Gateway is a native backend using the same API key as speech-to-text.
• The overlay is the control strip. v0.59.0 lets you choose which actions sit on the Windows overlay, rotate the speech language, and start a meeting from it. See SpeechKit framework.
• One Go package per speech provider. v0.64.0 splits every speech and realtime voice provider into its own SDK package, so an app that talks to one backend stops compiling the rest. Old names still work through v0.64 with a migration table and are removed in v0.65.0. The model catalog is now single-sourced, and entries carry a modality field.
• Owner-approved removal targets one node. StackKits v0.21.11 through v0.21.13 let owner-approved workload removal select one exact node placement from a verified multi-node Apply result, with absence evidence still correlated to the same Apply.
• Richer access manifest. v0.21.16 extends the native access manifest to v3 with application grouping, display names, lifecycle impact, and secret-free internal addresses for running services.

​

Bug fixes

• Cloud hardening can no longer lock you out. StackKits v0.21.10 preserves a non-root key-authenticated account before Cloud host-security disables root login, and v0.21.17 makes the hardened sshd policy win over provider and cloud-init defaults, so root key login stays disabled.
• Re-running Apply on ports 80 and 443 is safe. v0.21.18 lets an idempotent Apply reuse ports 80 and 443 only when Docker proves they belong to the current workspace’s Cloud core. Foreign or stale listeners still block.
• Workloads without a placement block just work. v0.21.11 applies the governed placement defaults when a StackSpec author selects a workload without a placement block. See Configure a stack spec.
• A failed Apply tells you what to do next. v0.22.0 ends a failed Product Apply with retry, status, and logs guidance plus the Run-ID, and the installers resume an existing workspace instead of dying.
• A refused voice session explains itself. SpeechKit v0.62.0 makes a refused Voice Agent session name the one thing to change, and every error carries the id support needs to find the request.
• Spoken numbers come out as digits. v0.61.0 writes spoken numbers as digits and strips German filler words like äh and ähm, including when the language is set to all languages.
• The assistant stops interrupting itself. v0.61.0 rejects the agent’s own voice as barge-in, stops a hung provider close from leaving a session stuck ending, and plays the reply at the correct 24 kHz rate.
• Windows shows the real version after an update. v0.61.0 makes Install and restart display the published version instead of the previous one.
• Full-capture Dictation respects your engine choice. v0.62.0 uses the selected speech engine instead of always trying Deepgram first.
• The keyboard behaves on Android. v0.60.0 puts SpeechKit on the Gboard suggestion strip and restores the provider buttons on the keyboard, and v0.61.0 keeps a tester connection working until you explicitly switch to a hosted connection.

​

Techstack open-source Alpha, SpeechKit on the Android keyboard, StackKits Cloud core

​

New features

• Techstack Windows Alpha, open source. The Techstack open-core client is now published under AGPL-3.0 at kombifyio/TechStack, and the unsigned Windows Alpha is available for local-server feedback. Techstack is the optional orchestration and operator experience on top of StackKits, and this Alpha supports the account-free local path. See Techstack overview, availability, and Install Techstack Alpha on Windows.
• Dictation and Voice Agent, one tap above the keys. SpeechKit v0.58.0 adds a row above the Android keyboard to dictate on device or through your SpeechKit server, or talk to Deepgram, AssemblyAI, or GPT. Actions that need a server say so instead of failing when pressed. See SpeechKit modes.
• SpeechKit ships its own Android keyboard. v0.57.1 replaces the borrowed IME with a SpeechKit-branded keyboard (a HeliBoard fork). Tap the voice key and the dictation panel opens inside the keyboard, where you were typing, without an app switch or a lost cursor. The complete corresponding source is published alongside the release.
• Ten first-class use cases in StackKits. v0.21.1 completes the CUE authorities for all ten use cases, including Vault, Media, and draft AI/Dev/Remote, plus post-1.0 draft Mail and Game. Existing Basement and Cloud Kit roles are bound to them. See Choosing a kit.
• Provider-neutral Cloud core runtime. v0.21.1 implements the provider-neutral Cloud core runtime with a guarded local Product Apply, live Verify observation, and the exact access-manifest v2 projection for base, auth, id, and coolify. Provider and credential custody stay in Techstack. See Cloud Kit.
• Configure a SpeechKit mode before you turn it on. v0.57.1 keeps Transcribe, Assist, and Voice Agent settings open and editable while the mode is switched off, and no longer throws you out of the page when you toggle a mode off. See SpeechKit framework.

​

Updates

• Desktop shell branded “SpeechKit UI”. v0.57.1 gives the desktop window the SpeechKit mark and title so it no longer identifies itself as a generic shell.
• foundation replaces the base schema package. StackKits v0.18.22 renames the shared CUE schema package from base to foundation and requires foundation/stackkit.cue in published archives. Product kits are unchanged and still extend #FoundationStackKit.
• Retired the base-kit installer argument. v0.18.22 rejects the retired base-kit slug instead of silently rewriting it to basement-kit. Use basement-kit. See Choose a kit.

​

Bug fixes

• Cloud Apply reaches the posture it declares on a stock Ubuntu host. StackKits v0.21.9 installs the packages backing Cloud host-security units and its ruleset instead of failing on absent nft and fail2ban, and only reloads a running sshd while ensuring the privilege-separation directory on socket-activated hosts.
• Cloud host-security failures now name the failing unit. v0.21.8 retains the bounded command and process output when Cloud host-security reconciliation fails, so a live target reports the exact failing unit.
• Exact Cloud core Verify handles the socket proxy and health-checked services. v0.21.6 accepts the not-configured observation for the intentionally healthcheck-less socket proxy, and v0.21.7 retains exact runtime-owned public-route health outcomes after validating them against Cloud core module probes.
• Cloud Kit Apply stops failing closed on dispatcher metadata. v0.21.4 accepts dispatcher plan-class metadata on Cloud Kit owners and host-admission so Product Apply no longer fails with health:stackkits-host-admission=executor-failed. v0.21.5 executes all-local Cloud Kit owners through their exact local router. See Review, plan, and apply.
• Verify and Status reject stale Cloud service projections. v0.21.1 rejects stale, foreign, or mutated Cloud service projections during Verify and Status readback.
• No more duplicate SpeechKit launcher icons on Android. v0.58.0 stops the keyboard fork from adding a second launcher that opened the keyboard’s settings instead of the app.
• The voice key is there when you install SpeechKit. v0.58.0 makes the voice key visible without enabling a second input method, and the dictation panel now opens over the keys instead of at the top of the screen.
• Voice Agent status is truthful on Android. v0.57.1 stops advertising the Voice Agent as unreleased and lets the SpeechKit keyboard actually appear, with the dictation panel mounted inside its input view.

​

StackKits v0.15.0 – v0.18.1: guided installer, service-controllable rollouts, and a provider-neutral Cloud core

​

SpeechKit v0.53.0 – v0.57.0: one voice assistant everywhere, Android Voice Agent, hosted web assistant

​

New features

• Guided installer to a running homelab. StackKits v0.17.0 ships a full guided installation with three modes and a v2 init flow that accepts --use-case, --platform, and --enable. The installer bootstraps Docker and storage roots directly, works safely under sudo, and prints truthful native credentials. See Quickstart and Choosing a kit.
• Provider-neutral Cloud core with public TLS routes. StackKits v0.15.0 resolves a required cloud-core workload for base, auth, id, and coolify, renders one digest-pinned Compose project, and executes it through a closed local runtime owner on an externally supplied host. v0.18.0 makes Cloud Kit own and materialize public TLS routes end to end. Cloud runtime custody stores only service-scoped PocketID, TinyAuth, and Coolify material. Provider credentials, server credentials, leases, and inventory stay outside StackKits. See Cloud Kit.
• stackkit service command for post-rollout control. v0.16.0 adds stackkit service start|stop|restart <service-key> and stackkit service logs <service-key> for StackKits-managed Cloud core services. Mutations require --owner-approve, and --json returns a versioned action result. Use it to bounce or inspect one service between full applies. See Day 2 operations.

stackkit service restart coolify --owner-approve --json
stackkit service logs auth --tail 100

• Full machine-readable output on apply, status, verify, and logs. v0.15.8 keeps stackkit apply --json, stackkit status --json, and stackkit verify --json parseable on success, denial, and failure. stackkit logs returns redacted, bounded pages with digest-bound cursors and explicit truncation metadata. The MCP connector exposes the same structured evidence. See MCP connector and Operate a StackKit.
• Architecture v2 runtime listeners for product-owned application delivery. v0.15.3 exposes runtime listeners on Architecture v2 catalog projections so applications can drive their own delivery without moving lifecycle authority out of StackKits. See Application delivery.
• Android Voice Agent client. SpeechKit v0.57.0 adds a real-time Voice Agent to the Android app. Press to talk, hear the answer, and watch the transcript build as you speak. The SpeechKit keyboard opens the assistant in place, so a conversation stays out of the text field while dictation still writes into it. See SpeechKit modes.
• Hosted /assistant web page on every SpeechKit server. v0.55.0 serves a ready-to-use Voice Assistant page. Open it in any browser, connect, and start talking. See SpeechKit overview.
• Voice Assistant appearance settings and one visual language everywhere. v0.55.0 adds Settings to pick the animation (Aura orb or Waveform), the centre logo, and whether the live transcript shows by default, with a live preview. v0.57.0 publishes the shared orb so the Android app, Windows overlay, and hosted web page draw the same assistant from a single source. See SpeechKit framework.
• Deepgram Flux v2 across all voice paths. v0.54.10 makes Deepgram Flux v2 available for realtime Voice Agent, speech-to-text, and text-to-speech.
• Local Box media lifecycle. v0.54.0 plays media through the local device agent on a paired Kombify Box, so the audio path stays on your own network.

​

Updates

• Cloud Kit offsite backup is now opt-in. v0.16.1 makes offsite backup an explicit choice on Cloud Kit. See Back up and restore.
• Traefik admin API stays localhost-only on Cloud Kit. v0.18.1 keeps the Traefik admin API bound to localhost in both product and public release contracts.
• Rebuilt Windows Voice Agent overlay. SpeechKit v0.56.0 turns the overlay into one dark, readable panel centred on the animation and the conversation. Controls appear only when you reach for them. Quick Note moved to the overlay, and every overlay icon has a tooltip.
• Silent updater with a distinct verification phase. v0.52.11 installs updates silently, v0.52.12 has SpeechKit relaunch itself after install, and v0.54.10 shows a distinct verification phase instead of appearing to stall at 100%.
• Numeric pre-1.0 releases are back on the public surfaces. v0.52.9 publishes numeric v0.x releases as normal GitHub releases and advances the canonical latest alias so the website, installer mirror, and in-app updater agree on the newest build. v0.52.10 restores the missing public release notes for v0.51.37 through v0.52.9.

​

Bug fixes

• Apply binds mutation to the expected plan hash. StackKits v0.17.0 prevents apply from running against a plan that changed underneath it. See Review, plan, and apply.
• Authenticated public runtime image pulls. v0.16.8 authenticates the public runtime image pulls that a fresh install performs.
• Held-to-talk release ends the conversation. SpeechKit v0.56.0 closes the dialog when hold-to-talk is released. A short window still lets an immediate second press pick up where you left off.
• Launching SpeechKit opens SpeechKit. v0.56.0 brings up the window from the taskbar or Start menu and returns to the running app on a repeat click.
• Android keyboard setup detects our own keyboard. v0.57.0 stops the setup flow from claiming a step is done before it is.
• Transcripts keep their own language. v0.54.0 no longer labels transcripts as German regardless of what was spoken, so custom vocabulary and replacements apply to the language you actually used. Silent-recording drops also stop: when a recording produces no text because the spoken language did not match the configured one, SpeechKit says so instead of dropping it.
• Server keeps running through a transient probe failure. v0.54.1 stops a single failed probe from killing the SpeechKit server, and unlatches the local Whisper ready flag when it recovers.
• Model setup retries transient Whisper downloads. v0.54.10 retries instead of failing the install on a transient network error.