The second half of the v4 series was about two things: making Xeref reachable from anywhere, and making its agents honest about what they can actually do. By v4.9 you can control your workspace from your phone, and every system agent answers "what skills do you have?" from real configuration instead of guesswork.
v4.9 — Remote Control
Control your Xeref workspace from your phone without installing anything. Pair your Telegram bot once from Settings → Telegram, then text commands to create tasks, chat with your agents, and capture ideas on the go. Prefer a browser? Open xeref.ai on any phone, sign in with the same account, and the full dashboard is ready.
v4.8 — Per-Agent Skills and a Real Skills Matrix
Per-agent skills: XerefClaw, Xeref Agents, and XerefHermes now each have their own skill set instead of sharing one toolbox. XerefClaw keeps the full 9 skills (tasks, projects, memory, documents, web), Xeref Agents focuses on research, and XerefHermes specialises in memory and knowledge. Ask any agent what it can do and the answer comes from real configuration.
Skills matrix: The Skills page (/customize/skills) now shows exactly which skills each system agent has, grouped by category, and the agent picker in chat displays a skill count for every agent.
How Xeref Works: The About page gained an interactive showcase that walks through the product story — Build, Orchestrate, Deploy — with a scroll-linked visual that follows you as you read.
Hardening: Every push now runs an automated secret scan across the full git history, dependency vulnerability alerts and automated security fixes are enabled, and CI enforces a code-coverage floor. The GitHub connection also shows which organizations your token can reach.
v4.7 — Text Transforms and MCP Server Settings
Text transforms in the chat composer: Press Ctrl+. to polish what you wrote into a clearer message that keeps your voice, or Ctrl+Shift+. to turn rough notes into a structured, AI-ready prompt with role, task, and context sections. Transforms apply to just the selected text, or the whole message when nothing is selected.
MCP Servers settings (Xeref+): Agent Manager settings gained an MCP Servers tab — every server in the catalog with its description and tool count, a details pane showing exactly which tools each server exposes, and one-click enable/disable. Two new servers joined the catalog: Notebooks and Visualization (Apache ECharts rendering).
v4.6 — Claude Fable 5 on Ultra
Anthropic's new Mythos-class model is now available on Ultra. Select Fable 5 in the chat model picker for the deepest planning and long-horizon reasoning on the platform, with a 1M token context. Opus Plan Mode routes planning-shaped queries to Fable 5 automatically, and the Plans generator uses it for Ultra accounts. Plan generation also got more resilient: if the primary model declines, it automatically retries on Claude Sonnet 4.6 instead of failing.
v4.5 — Builder Catalog: 22 → 48 Capabilities
The /builder (XerefClaw) catalog now offers 48 agent capabilities across all six CLAWS categories — new entries range from SMS via Twilio and Microsoft Teams (Connect) to OCR and video understanding (Listen), knowledge graphs and hybrid search (Archive), Notion, Stripe, and Linear integrations (Wire), RSS and website-change monitors (Sense), and agent-architecture patterns like ReAct loops, guardrails, and structured output.
v4.4 — The Download Page
xeref.ai/download is now a real page explaining how to continue your workspace on any device: a three-step walkthrough, a list of everything that travels with your account (projects, agent chats, Pinecone memory), and an honest roadmap card for the native mobile app.
Together, these releases close out v4 with a platform you can carry in your pocket and agents whose capabilities are inspectable, not implied.