Directories
¶
| Path | Synopsis |
|---|---|
|
Package browser is unblink's orchestrator.
|
Package browser is unblink's orchestrator. |
|
Package content converts non-HTML page bodies — JSON, plain text, RSS/Atom/JSON feeds, PDF, and images — into Markdown for the read pipeline.
|
Package content converts non-HTML page bodies — JSON, plain text, RSS/Atom/JSON feeds, PDF, and images — into Markdown for the read pipeline. |
|
feed
Package feed converts RSS, Atom, and JSON feeds into a Markdown summary.
|
Package feed converts RSS, Atom, and JSON feeds into a Markdown summary. |
|
image
Package image produces a Markdown manifest for an image response — type, dimensions, byte size, and filename — without inlining the pixels.
|
Package image produces a Markdown manifest for an image response — type, dimensions, byte size, and filename — without inlining the pixels. |
|
pdf
Package pdf extracts text from PDF documents and renders it as Markdown.
|
Package pdf extracts text from PDF documents and renders it as Markdown. |
|
Package dom is the only package that imports golang.org/x/net/html (and, in later phases, goquery/cascadia).
|
Package dom is the only package that imports golang.org/x/net/html (and, in later phases, goquery/cascadia). |
|
Package emit serializes reduced content into the representation an AI agent consumes.
|
Package emit serializes reduced content into the representation an AI agent consumes. |
|
Package fetch is unblink's HTTP-client-as-browser.
|
Package fetch is unblink's HTTP-client-as-browser. |
|
Package js is unblink's quarantined JavaScript engine.
|
Package js is unblink's quarantined JavaScript engine. |
|
Package mcpserver exposes the browser as an MCP server.
|
Package mcpserver exposes the browser as an MCP server. |
|
Package page defines the core data model that flows through the unblink pipeline.
|
Package page defines the core data model that flows through the unblink pipeline. |
|
Package ratelimit provides a process-global, per-host request rate limiter so unblink throttles requests to each host across all its HTTP clients.
|
Package ratelimit provides a process-global, per-host request rate limiter so unblink throttles requests to each host across all its HTTP clients. |
|
Package reduce performs unblink's semantic reduction: it strips visual-only junk (scripts, styles, nav, ads, hidden elements) and extracts the meaningful content of a page.
|
Package reduce performs unblink's semantic reduction: it strips visual-only junk (scripts, styles, nav, ads, hidden elements) and extracts the meaningful content of a page. |
|
Package robots parses robots.txt (the Robots Exclusion Protocol) for the purpose of *exposing* a site's crawl rules to an agent — not enforcing them.
|
Package robots parses robots.txt (the Robots Exclusion Protocol) for the purpose of *exposing* a site's crawl rules to an agent — not enforcing them. |
|
Package search adds an optional web-search entry point behind a small Provider interface, so unblink stays fully self-contained by default and only reaches an external service when the operator configures one.
|
Package search adds an optional web-search entry point behind a small Provider interface, so unblink stays fully self-contained by default and only reaches an external service when the operator configures one. |
|
Package session holds unblink's per-session state: a cookie jar (via a dedicated fetch.Client), a navigation history, and an optional persistent JavaScript runtime ("live context") bound to the current page.
|
Package session holds unblink's per-session state: a cookie jar (via a dedicated fetch.Client), a navigation history, and an optional persistent JavaScript runtime ("live context") bound to the current page. |
|
Package sitemap parses the sitemaps.org XML formats — a <urlset> of page URLs or a <sitemapindex> of child sitemaps — using only the standard library.
|
Package sitemap parses the sitemaps.org XML formats — a <urlset> of page URLs or a <sitemapindex> of child sitemaps — using only the standard library. |
|
Package tokens provides cheap token estimation and Markdown pagination so the browser never returns an unbounded blob to the model.
|
Package tokens provides cheap token estimation and Markdown pagination so the browser never returns an unbounded blob to the model. |
|
Package webext models Firefox/Chrome WebExtensions well enough to load a user-supplied extension (uBlock Origin and friends) at runtime.
|
Package webext models Firefox/Chrome WebExtensions well enough to load a user-supplied extension (uBlock Origin and friends) at runtime. |
Click to show internal directories.
Click to hide internal directories.