Skip to main content

🧚 GLM 5.3 lands, Mobile Peering, and alerts go platform-wide

Released on August 31, 2026

The GLM 5.3 family lands officially — first teased as coming soon, now available across standard, priority and fast tiers — alongside Mobile Peering so you can drive a local agent on your desktop right from your phone, alerts becoming a first-class citizen across the whole developer platform, new email and plugin powers for TF Code, and a smoother, more reliable experience across the board! 🚀

🤖 New Models​

GLM 5.3 is here​

The GLM 5.3 family is now available across standard, priority and fast tiers. Built on Fireworks with an 800,000-token context, GLM 5.3 is a reasoning model with tool calling and planner support, ready for long-horizon, agentic work.

Key Features:

  • Three tiers - GLM 5.3, GLM 5.3 Priority and GLM 5.3 Flash for standard, priority and fast workloads
  • Huge context - 800K context length for long documents and extended conversations
  • Reasoning and tools - a reasoner with tool calling and planner availability out of the box

🖥️ Mobile Peering — Your Desktop, in Your Pocket​

Drive your local agent from your phone​

With Mobile Peering, you can now chat with and drive a local agent running on your desktop computer straight from the ToothFairy mobile app — whether you are on the same network or halfway across the world. Pair your phone with a one-time code, and your desktop appears in the app with a live online status, ready to chat.

Key Features:

  • Relay mode (from anywhere) - connect to your local agent over the internet, with no need to be on the same Wi-Fi
  • One-time pairing code - pair your phone in seconds and see your desktop's live online status at a glance
  • Remote chat - chat with your local agent from your phone, with sleep protection keeping your machine awake and reachable

Note: Mobile Peering is in beta and rolling out now — it will be made available soon to all users and regions. To try it today, enable experimental features in the mobile app and look for the Mobile peering section on each side.

🔌 Alerts Across the Platform​

Create and manage alerts everywhere you build​

Alerts are now a first-class citizen across the entire developer platform. An alert notifies your chosen recipients — in-app, push, email or webhook — the moment all of its conditions match, and you can now create and manage them over the REST API, from both SDKs and both CLIs, and through MCP.

Key Features:

  • Full alert lifecycle - create, get, update, list, delete, plus pause and resume, across the REST API, SDKs, CLIs and MCP
  • Condition builder - field, operator and value conditions that all must match, with three trigger types — analytics, system and custom
  • Flexible notifications - choose recipients and channels (in-app, push, email, webhook), set severity and a cooldown between repeats

One manager, every operation — SDKs and CLIs​

Both the Python and Node SDKs ship a new alerts manager — client.alerts — with the full operation set, typed alert models and automatically parsed conditions. Both command-line interfaces gain a matching command set — create-alert, list-alerts, get-alert, update-alert, delete-alert, pause-alert and resume-alert — with comma-separated recipients and a JSON passthrough flag so every newly available field is immediately usable from the terminal.

MCP gains alerts, triggers and predictions​

The ToothFairyAI MCP server learns three new tricks for your AI clients: create, get, update, delete and list alerts; the full trigger toolkit over MCP; and a predictions tool to run an OpenAI-compatible model prediction and receive the complete response text.

🎯 TF Code​

Send email and get plugin recommendations​

TF Code gains two new capabilities: your local agents can now send email notifications straight from a session, and TF Code will recommend the right plugins to switch on for your workflow.

What's New:

  • Email sending - trigger branded email notifications from your local agent, complete with a subject, body and optional call-to-action
  • Plugin recommendations - ask TF Code which plugins are worth enabling and get a tailored recommendation for your workspace

🛠️ Improvements & Fixes​

Reliability and polish​

  • Mobile workspaces - switch between workspaces and keep an eye on analytics, members and billing right from the mobile app
  • Widget tool-call details - embeddable chat widgets can now show tool-call details inline, so people using your widget see exactly what the agent is doing
  • Better API for automation - smoother, more predictable API hooks to power your backend automations and integrations
  • Deep Search reliability - bounded the search wait so a hung page can never stall a Deep Search run
  • Preserved thinking - reasoning now carries cleanly through API calls and internet search, with stray payload keys stripped to avoid provider errors
  • Consumption tracking - clearer tracking of consumption across backend services
  • Billing cap - new behind-the-scenes management of a billing cap
  • Mobile app 1.0.6 - refreshed App Store and Play Store listings with new dashboard and members screenshots
  • Faster signup - a smoother, quicker sign-up flow

📋 Summary​

This update includes:

  • GLM 5.3 - the GLM 5.3 family arrives across standard, priority and flash tiers, with 800K context, reasoning and tool calling
  • Mobile Peering - drive a local agent on your desktop from your phone, over the internet or your LAN, with one-time pairing codes and sleep protection
  • Alerts across the platform - full alert lifecycle over the REST API, SDKs, CLIs and MCP, with conditions, severities, cooldowns and per-channel notifications
  • TF Code - send email notifications and get plugin recommendations from your local agents
  • Reliability - mobile workspaces and widget tool-call details, a better API for automation, Deep Search and preserved-thinking fixes, clearer consumption tracking, a refreshed mobile app and faster signup

Thank you for using ToothFairy! We're constantly working to improve your experience. If you have any feedback or encounter any issues, please don't hesitate to reach out to our support team at support@toothfairyai.com.

Keep building magic with AI! 🧚‍♀️✨

v0.704.0

Last updated on