Skip to release notes
All updates
v0.9.0

Fostera v0.9.0 — Quick chat, audio & live web work

New Features

  • Added Quick chat: start a full-capability conversation without using or creating Soul memories. It automatically expires 24 hours after it begins, or you can save it to keep the complete chat and its files.
  • Added audio attachments in regular and Quick chats. Attach MP3, WAV, or M4A recordings up to 10 MB and 30 minutes long, play them back, and ask your Soul about what was said.
  • Added a live browser for web work. Watch your Soul navigate a page, then take over whenever a site needs you to drive.
  • Added the ability to attach a file from an earlier conversation instead of uploading it again.
  • Added Fostera, a ready-to-use default Soul that gives every new account a place to begin.
  • Added new model choices, including GPT-5.4 Pro, GPT-5.5 Pro, Grok 4.7, Qwen 3.8, DeepSeek V4.1 Flash, GLM 5.3, and Kimi K3.

Improvements

  • Improved chat responsiveness while typing or pasting in long conversations.
  • Improved Quick chat controls: files and abilities are easier to find, the header shows its deletion countdown from the start, and saving preserves your draft, position, and focus.
  • Improved Quick chat visibility with clear labels and Quick / Regular filters on Home and in the chat library.
  • Improved chat cards across Home, the chat library, Soul activity, and starred chats with consistent chat types, activity, location, and Quick-chat deletion status.
  • Improved model selection and Auto routing so each provider's input support, response limits, thinking settings, and live progress behavior are applied accurately.
  • Improved Synthesis with reusable chat settings, separate effort controls for contributors and the final writer, clearer model-team details, and more dependable recovery.
  • Improved on-device chat so selected local models and downloaded runtimes remain available after a reload or retry.
  • Improved web research so Souls can read more links, keep a live browser open for you to return to, and show clearer details for web work in a reply.
  • Updated prices for Premium 10x and 20x plans and for the 20M and 40M Resonance packs.

Bug Fixes

  • Fixed Quick chat isolation so it does not load or change Soul memories or other conversations before you save it.
  • Fixed proactive check-ins so they only send a verified message grounded in the right conversation; unsuitable or unavailable messages are now skipped instead of replaced with a generic prompt.
  • Fixed interrupted Synthesis and large-document reviews so they recover more reliably without losing completed work or misreporting Resonance.
  • Fixed several provider fallback and routing cases that could leave an otherwise answerable chat unfinished.
  • Fixed the update prompt appearing after every deployment.
  • Fixed attachment, project-file, and completed-chat cleanup cases that could leave a chat or its files in the wrong state.