Changelog

What we ship.

Every release in order. New features, improvements, fixes, and security updates, public and dated.

  1. v4.2

    Brand Voice 2.0, multilingual chatbots, 8 new models

    A new Voice editor with sliders, native multilingual replies for chatbots, and eight new models, including GPT-5 turbo and Claude Opus 4.7.

    • NewBrand Voice 2.0: train in 60 seconds with slider-based tone tuning.
    • NewMultilingual chatbot replies: auto-detect visitor language across 32 locales.
    • NewEight new models: GPT-5 turbo, Claude Opus 4.7, Gemini 2.5 Pro, Llama 3.3 70B, plus four open-weights.
    • Improved32% faster cold-start on chatbot embeds (≤ 14kb gzipped).
    • ImprovedWorkflows: per-block retry policies now configurable in the visual editor.
    • FixedDoc presence avatars sometimes flickered after the editor regained focus.
    • SecurityRotated all OAuth client secrets; no action required.
  2. v4.1

    Workflows GA, audit log export

    Workflows is generally available with human approval gates and webhook triggers. Plus Enterprise-grade audit-log export.

    • NewWorkflows GA: visual builder, webhook + cron triggers, branching, approvals.
    • NewAudit log streaming to S3, Datadog, and Splunk.
    • ImprovedAI Writer SEO checklist now flags missing internal links.
    • FixedWebhook signing header was occasionally truncated above 8kb payloads.
  3. v4.0

    A new workspace, and a Series B

    Re-launched workspace with a unified inbox for chat, docs, and workflows. Coincides with our $42M Series B.

    • NewUnified workspace combining chat, documents, and workflows.
    • NewReal-time presence and conflict-free editing in Documents.
    • New65+ models available out of the box; default routing by cost or quality.
    • ImprovedPricing: free tier expanded from 50 to 250 messages/month.
    • FixedImage Studio occasionally lost the prompt history after a 30+ minute idle.
  4. v3.18

    Voiceover library expansion + custom voices

    Sixty-plus languages and 200+ voices in Voiceover. Studio plan customers can clone custom voices.

    • NewVoiceover: 60+ languages, 200+ voices, two custom-voice slots on Studio.
    • ImprovedDocument import: .docx fidelity for footnotes and tracked changes.
    • FixedMarkdown export from Documents stripped some inline code styles.
  5. v3.17

    EU-only data residency for all plans

    You can now mark a workspace as EU-only. All inference, storage, and logs stay in Frankfurt.

    • NewEU-only data residency toggle (per workspace).
    • NewRegion-aware provider routing for OpenAI, Anthropic, Google models.
    • ImprovedReduced p95 latency on EU inference by 38%.
  6. v3.16

    AI Code in beta, 35 languages

    Generate, explain, and debug code across 35 languages with citations and brand-aware comments.

    • NewAI Code beta: 35 languages, citations, repo-aware suggestions.
    • ImprovedChatbots: KB sync now incremental, under 30s for typical updates.
    • FixedBrand Voice: sliders occasionally snapped to defaults on Safari.

What's next

Want updates the moment they land?

Sign in to see your workspace's update log, or get in touch about the roadmap.

Sign inContact us