What changed in Player One

Platform and admin updates from 8 to 12 June 2026, newest first. See the latest updates.

Major release

Smoother voice, cleaner branding and more model control

This update brings a big round of voice-call polish, finishes off the Player One branding and version cleanup, and gives you more visibility and control over models. Sidebar, search and WhatsApp setup all got steadier too.

Improvement
Smoother voice calls
Model changes during a call now take effect straightaway, short spoken commands are recognised more reliably, and calls feel snappier when stopping or handing turns back, with fixes for warm-up, audio crackle and hang-up handling.
Improvement
Consistent Player One branding
Player One now speaks more consistently as Player One, with older naming cleaned up, and the platform version display simplified into one clear release instead of separate moving parts.
New
Per-client model usage
The gateway can now track model usage per client, with a new top-bar indicator, plus a new models tool that makes it easier to list, search and switch models from inside Player One.
Improvement
Friendlier model picker order
Preferred Claude options now appear in a more useful order in the model picker.
Improvement
Sidebar and chat list polish
Pinned areas show clearer icons, agent names are easier to spot, recent chat groups have more readable labels, and chat lists no longer use internal system messages to build titles or previews.
Improvement
Cleaner agent organisation
Agents moved to a cleaner per-agent file structure by default, with stronger protection against simultaneous config writes during agent creation.
Improvement
Clearer search results
Context search results are easier to read when several matches come from the same file, and Fold guidance was reorganised so agents pick the right search behaviour more reliably.
Improvement
Steadier WhatsApp linking
WhatsApp linking received several fixes focused on avoiding device revocation after pairing, recognising pairing mismatches sooner, and clearer self-test guidance.
Improvement
Faster, clearer startup
Gateway builds are faster and more predictable, startup logs more useful boot phases, and the platform fails clearly instead of continuing in a broken state.
Fix
Quality fixes
Clearer error messages from model calls, safer channel handling, automations gated off by default, more useful debug session views, and the mobile screen now stays awake during an active chat run or voice call.
Major release

Clearer chat activity, better job control and smoother previews

This update makes Player One easier to follow while work is happening. Chats now show clearer activity cards, jobs pause more politely when you step in, model and agent pickers are easier to read, search results are more useful, and previews reopen and recover more reliably.

Improvement
Clearer chat timeline
System events, model changes and interruptions now appear as proper timeline cards, making conversations easier to follow.
Improvement
Model switches are visible
When the model changes inside a conversation, Player One now shows a clear divider instead of making the thread feel like it jumped.
Improvement
Cleaner cancellation
Job cancellation controls and markers are clearer, so it is easier to stop work cleanly and see what happened afterwards.
Improvement
Better agent picker
Agent names and descriptions now appear together, making it easier to choose the right helper without opening each one first.
New
Claude Fable 5 added
Claude Fable 5 has been added to the curated model list, alongside friendlier model names and better fallback syncing.
Improvement
More useful search results
Search and context results now return better snippets, with results centred around the actual match and clearer agent descriptions.
Improvement
Lighter indexing
The search indexer now uses less memory, helping the system stay responsive during reindexing and ongoing updates.
Improvement
Smoother tool loading
Tool-loading continuations are more reliable, helping agents keep working after new capabilities are attached.
Improvement
Better previews
Preview panes reopen when asked, close cleanly when expired, and report completed work more consistently.
Improvement
Account identity fixes
Re-authentication, account removal and live identity checks now behave more reliably after model switches.
Major release

Faster start-ups, cleaner skills and friendlier recovery

This release makes Player One start more steadily, gives agents a cleaner skills system, improves the agent-creation flow, and makes reconnecting or logging back in feel less alarming. It also tightens model handling, chat polish and restart guidance.

Improvement
Steadier start-ups
Player One now starts more calmly and uses memory more carefully while it boots, especially on smaller hosted environments.
Improvement
Better boot visibility
Improved internal visibility around start-up time and memory use makes slow starts easier to spot and improve.
Improvement
Skills are easier to browse
The Skills area now has clearer separation between system and custom skills, paging for long lists, and previews before opening.
Improvement
Smoother agent creation
Creating a new agent now gathers the important details up front and then builds the agent in one go, instead of stopping through lots of small questions.
Improvement
Clearer reconnection
Gateway reconnection messages are calmer and clearer, with a more consistent overlay while the app reconnects.
Improvement
Login recovery button
When a session needs refreshing, Player One now gives you a clearer login button rather than leaving the app in an uncertain state.
Improvement
Predictable model lists
Model information shows more reliably after cold starts, and curated model lists are more predictable.
New
Grok indicator
A Grok indicator now appears alongside Claude, ChatGPT and Ollama in the top bar.
Improvement
Better chat behaviour while composing
Chat now does a better job of keeping the thread pulled to the newest message while you are composing.
Improvement
Clearer provider overloads
Provider overload errors are labelled more accurately, so temporary provider issues are not confused with normal rate limits.