Patch release.
Fixed
- MCP:
xberg mcpextract/extract_batchstructured output no longer fails its own declared output schema — optionalerrors/crawl_*fields are now correctly optional, so MCP clients (e.g. Claude Code) accept normal extractions. (#1372) - CLI:
install.shno longer creates a self-referentialxbergsymlink loop, and selects the glibc build on standard Linux distros instead of always downloading the musl build that failed to run on glibc systems like Ubuntu. (#1371) - CSV: all-text first rows are treated as headers instead of rendering a broken blank header row. (#1369)
- Docs: search input is no longer hidden by an oversized clear icon (docs-theme 0.4.1). (#1370)
Internal
- Added integration test coverage for the OpenWebUI-compatible extraction endpoints (
PUT /process,POST /v1/convert/file).
Zig
Add to your build.zig.zon:
.dependencies = .{
.xberg-zig = .{\n .url = \"https://github.com/xberg-io/xberg/releases/download/v1.0.12/xberg-zig-v1.0.12.tar.gz\",\n .hash = \"xberg-1.0.12-iV1GrjGrGRep67-R7NzIkCA0PazIIadjgcP_7HXULMSk\",\n },\n},\n```\n