Canoryn 0.2.1

Fixes for problems in 0.2.0 that blocked everyday use — chat and voice acting again, browser workflows running, and a smoother CLI install.

Fixed

• Asking Canoryn to do something works again — requests like “build me a workflow” replied in text without actually doing the work. Chat, Architect chat beside the canvas, and voice were all affected.
• Workflows with a browser step run again — after switching a browser profile, watching an element, or picking something on a page, the whole workflow refused to run and reported an unknown setting.
• Installing the command-line tool no longer asks for your password — it installs into your own account. If that location isn’t on your PATH yet, Canoryn offers to add it and shows you the exact line first.
• Chat no longer fails on some OpenAI models when a step needs tools; those requests were rejected before reaching the model.

Install

Canoryn is still in beta and is not notarized yet, so macOS may show a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings → Privacy & Security → scroll down → Open Anyway

Thanks for trying the beta — support@procodeai.com.

Canoryn 0.2.0

Browser nodes that act like a real browser desk, terminals agents can actually read, models you pick on purpose, and a clearer canvas for building and running workflows.

Added

• Browser that behaves like a browser — search a phrase from the address bar; localhost stays http for local dev servers; idle pages sleep after about 30 minutes and come back where you left them.
• Point at a thing, hand it to an agent — click an element on a page, add a note about why, and wire it into an AI step (including multi-element picks).
• Tell me when this changes — watch an element and let a workflow run on price drops, build status, stock arriving, and similar signals.
• Capture a region for vision — for charts, PDFs, and anything the page structure can’t describe.
• Terminal steps agents can read — live terminal on the canvas; an agent can run a command and see whether it passed. The shell can outlive the view.
• Pick your models — star the few you use; those are the ones in the picker. Badges show which support tools, vision, and JSON.
• Attach images and files in chat — several at once, including PDFs and text, through one multimodal path.
• Clearer nodes and run insight — redesigned node chrome, execution context, per-node logs, and run history you can search with Run IDs.
• Secrets where they belong — environment and Keychain-backed secrets for workflows that need credentials without stuffing them into the graph.
• Richer authoring for agents — YAML source surface, canvas query language, and inspect formats (including mermaid/YAML) so CLI/MCP agents can read and edit graphs more precisely.
• Media control on the canvas — find and silence background audio from browser sessions.
• MCP run inspection — list runs and inspect steps from external agents without opening the UI.

Changed

• Model selection is literal — no more provider tiers quietly choosing a different model than the one you picked.
• Zooming no longer changes the document — wires stay attached to their ports at every zoom level; geometry stays scale-invariant.
• Downloaded / restored workflows land disarmed — they stay off until you enable them; installs are size-capped and checksum-verified.
• Version history is durable — more than three snapshots, and failures are reported instead of swallowed.
• Canvas LOD and chrome — smoother detail at distance, animated active wires, and a steadier grid while you work.

Fixed

• Workflows saved without their .cryn extension on some builds — Finder preview and double-click open work again.
• Chat failing outright on some models when tools were attached.
• Browser resume flicker, freeze-on-switch, and pages loading at the wrong size.
• A page you’re looking at is never the idle-eviction victim.
• Wire/port meeting points and low-detail node readability.
• Terminal focus, login-shell startup, and honest store/HTTP error reporting.
• Graph runner execution limits and cleaner LocalizedError messaging.

Install

Canoryn is still in beta and is not notarized yet, so macOS may show a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings → Privacy & Security → scroll down → Open Anyway

Thanks for trying the beta — support@procodeai.com.

Canoryn 0.1.6

Build workflows with chat or an external AI agent, run research boards, and trust enable/run behavior.

Added

• Canvas Architect Copilot — describe what you want in chat next to the canvas; Canoryn builds and edits the workflow instead of running the task immediately.
• Ask before it assumes — when something is unclear (location, schedule, missing tool), the copilot shows choices for you to pick.
• Edit without rebuilding — change one setting, swap a step, add or remove a node, or reconnect steps while keeping the rest of the workflow intact.
• External agents can author too — Codex, Claude, Cursor, and Antigravity can create, inspect, edit, enable, and run workflows through Canoryn’s MCP/CLI (no hand-editing project files).
• See what you can build — browse the live list of steps you can put on a workflow (with their inputs/outputs) before inventing something new.
• Dispatch other coding agents from a workflow — a canvas step can hand work to Codex, Claude, Cursor, or Antigravity on a repo and continue from the result.
• Research boards that actually run — compare sites in parallel, synthesize a report, and optionally email it — as a real workflow, not a static diagram.
• Safer custom tools — when Canoryn proposes a script tool, you see the full code before approving; approved tools show up as normal canvas steps with the source visible.
• Shorter workflow IDs in the CLI — short readable handles instead of full UUIDs.
• CLI & MCP settings — install and diagnose the command-line / agent connection from Settings; optional auto-allow for routine external runs (enable still asks).

Changed

• “Build a workflow…” means build — everyday shortcuts no longer fire the underlying action when you asked to design a workflow.
• Enable means it can run for agents/CLI — turning a workflow on publishes it so scheduled and external runs work; Play on the canvas still works while drafting.
• Clearer split — tools that edit a workflow stay separate from steps that live on the canvas, so agents don’t confuse the two.

Fixed

• Enabling a draft no longer leaves it “on” in the UI but blocked when an agent or CLI tries to run it; errors say why.
• Canvas Play works for scheduled / event-triggered workflows and no longer gets stuck on extra consent.
• Markdown reports keep their content after Play; browser pages survive collapsing groups.
• Custom tools you already trusted show up reliably as canvas steps after restart.
• Workspace and canvas focus/pan quirks that made editing feel sticky or unpredictable.

Install

Canoryn is still in beta and is not notarized yet, so macOS may show a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings -> Privacy & Security -> scroll down -> Open Anyway

Thanks for trying the beta - support@procodeai.com.

Canoryn 0.1.5

Release automation and CLI polish.

Added

• Release publishing now refreshes the public changelog feed and dispatches a docs rebuild after a successful release.
• Music planning has additional reflex evaluation coverage for safer routing changes.

Fixed

• Release feed publishing now preserves the latest stable release when older versions are backfilled.
• Sparkle update feed URLs are validated before the expensive archive step, so a bad feed configuration fails early.
• CLI stdout is cleaner for scripted runs and automation output.

Install

Canoryn is still in beta and is not notarized yet, so macOS may show a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings -> Privacy & Security -> scroll down -> Open Anyway

Thanks for trying the beta - support@procodeai.com.

Canoryn 0.1.4

Music that just works — for everyone.

Added

• Universal media controls. "Pause", "next song", "previous track" now control whatever is playing — Spotify, Apple Music, even a browser tab — using system media keys. No login or subscription needed.
• Smarter "play music." Canoryn now picks the best available way to play: your Spotify account if connected, the Spotify desktop app if installed (free accounts work), or YouTube Music in your browser as a free fallback that works for everyone. Saying "on Apple Music" or "on Spotify" always respects your choice.
• Play Spotify links directly. Give Canoryn a Spotify track/playlist link and it plays it on the Spotify desktop app — no Spotify account connection required.

Fixed

• No more false "ISSUES" warnings. Music commands (and things like "play the Game of Thrones theme") no longer show incorrect execution-trace errors after succeeding.
• Music requests no longer fail silently when a music service is turned off in Connections — Canoryn routes to one that's available instead.
• Web search results now include readable page content more often.

Changed

• Wallpaper fetching follows Unsplash's API guidelines (download attribution + photographer credit).

Install

Canoryn isn't notarized yet, so macOS shows a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings → Privacy & Security → scroll down → Open Anyway

Thanks for trying the beta — support@procodeai.com.

Canoryn 0.1.3

Fixed

• Music works without Spotify. "Play music" and mood-based ambiance now route to Apple Music automatically in builds without Spotify configured — mood music no longer silently does nothing. (When Spotify is configured, it's still used.)

Changed

• Unsplash wallpaper compliance. When Canoryn fetches a desktop wallpaper from Unsplash it now follows Unsplash's API guidelines (download-attribution endpoint + photographer credit captured).

Install

Canoryn isn't notarized yet, so macOS shows a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings → Privacy & Security → scroll down → Open Anyway

Thanks for trying the beta — support@procodeai.com.

Canoryn 0.1.2

This one makes Canoryn able to update itself.

Fixed

• Automatic updates now work. Earlier builds couldn't detect new versions (the update feed was returning an error). Canoryn now checks for updates on launch and installs them cleanly.

Added

• Update indicator in the sidebar. When a new version is ready, a tappable "Update available" pill appears at the top of the sidebar — one click downloads and installs it.

Note for v0.1.0 / v0.1.1 users

Those builds shipped with the broken update feed, so this is the release that gets you onto working auto-updates. If your app doesn't prompt you automatically, grab v0.1.2 here once — every update after this is automatic.

Install

Canoryn isn't notarized yet, so macOS shows a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings → Privacy & Security → scroll down → Open Anyway

Thanks for trying the beta — support@procodeai.com.

Canoryn 0.1.1

A quick, important patch.

Fixed

• No more crash on "Connect Spotify." Builds without Spotify configured no longer crash when the Spotify connection is opened from Settings.

Changed

• Spotify is temporarily unavailable in public builds. We're aligning with Spotify's Developer Policy, which restricts Web-API control for voice assistants. Apple Music control still works today, and broader system media controls + Apple Music music support is coming in the next release.

Install

Canoryn isn't notarized yet, so macOS shows a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings → Privacy & Security → scroll down → Open Anyway

Thanks for trying the beta — support@procodeai.com.

Canoryn 0.1.0 — first public beta 🎉

Canoryn is a local-first AI agent builder for macOS. Build private AI agents that run on your models — local or cloud — and act on your actual Mac. Native, offline-capable, and yours.

Highlights

• 🏗️ The Architect — a visual canvas to build agents: wire triggers, AI nodes, and native macOS actions into agents you run, save, and reuse. Like a game-engine Blueprint, but for AI agents.
• 🎙️ Talk or type — a menu-bar assistant with a customizable shortcut; speak or chat, and it acts on your Mac.
• 🧠 Bring your own models — fully private, offline inference via Ollama (Llama 3.2, Qwen 2.5…), or plug in your own OpenAI / Anthropic / Gemini / Groq keys. You own the brain.
• 🔌 Native integrations — control Finder, Calendar, Reminders, Apple Music / Spotify, system settings, and more.
• 📄 Portable .cryn agents — every agent is a readable file you own, version-control, and share.
• 🔒 Local-first & private — a real native macOS app; your data stays on your machine.

This is an early beta

0.1.0 is our first real public milestone — it's evolving fast and there will be rough edges. Tip: for the most reliable agents on local models, pin the tools on your AI node (or bring a frontier cloud key for open-ended tasks). We'd love your feedback.

Requirements

• macOS 14 or later (Apple Silicon & Intel)
• ~20 MB download (native app — no Electron)

Installing

Canoryn isn't notarized yet (no paid Apple Developer account), so macOS shows a one-time warning. To open it:

• Easiest (Terminal): xattr -dr com.apple.quarantine /Applications/Canoryn.app
• Or: System Settings → Privacy & Security → scroll down → Open Anyway

What's next

Reliability hardening, a memory that feels like a mind, more native actions, and a blueprint gallery.

Found a bug or have an idea? support@procodeai.com