v1.0.101
Released: 2026-04-05
Added
- Mobile Layout — Responsive design for mobile devices (
d504661) - Agent Time Awareness — Current datetime injected into LLM context for accurate time-related answers (
4a870ca) - GitHub MCP Server — Pre-configured GitHub MCP server, works out of the box (
0b03510) - Ollama Auto-detection — Auto-discover local Ollama on startup with pre-configured models (
8bb95fa) - Model Grouping — Model providers grouped by local/cloud with section headers, local-first sorting (
2210df0,5cb3981) - Zhipu AI Upgrade — Updated models to GLM-5-Turbo / GLM-5V-Turbo (
4b5991d) - Chat Suggestion Cards — Updated to showcase core features (memory, tools, multi-agent, skills) (
9ae3b51) - VitePress Docs — Versioned documentation, bilingual zh/en, release workflow guide (
e62e36f,3fb3506)
Fixed
- SPA Route Refresh 401 — Frontend route refresh no longer returns unauthorized (
c992921) - i18n Window Title — Window title now reads from language pack dynamically (
f256dcd) - Tool Guard Rule Names — Aligned runtime
@Toolmethod names with guard rule names, added audit config (f4ede77) - Zhipu Connection Test 404 — Switched to chat completions endpoint instead of
/v1/models(290d1d8) - Suggestion Card Misalignment — Widened grid to prevent misalignment with longer text (
f4a62fb)
Changed
- SpringDoc OpenAPI — Replaced Knife4j with SpringDoc 2.8.16 (
03887a9) - Copyright year updated to 2026 (
935cc0b) - Frontend version read from
package.jsondynamically (935cc0b)