PullMD: Self-hosted URL-to-Markdown MCP server for cleaner LLM web reading

PullMD is a self-hosted Docker stack that converts web pages into clean Markdown before sending them to MCP-compatible AI tools like Claude Code. For indie developers, the value is lower token waste, better mobile-to-AI workflows, and more control over privacy and infrastructure, though setup complexity and browser-side dependencies may limit casual adoption.

[ KEY POINTS ]

Targets a real pain point: raw HTML wastes context and tokens when AI agents read web pages.
Designed for MCP workflows with simple integration for Claude Code, Claude Desktop, Cursor, and similar clients.
Uses a multi-stage extraction strategy, including Markdown endpoints, readability parsers, and optional Playwright fallback.
Self-hosting is attractive for privacy-conscious indie developers, but operating a three-service stack adds maintenance overhead.
Most relevant as a practical productivity tool rather than a major platform shift or model-level change.

Originalwww.reddit.com/r/ClaudeAI/comments/1sxzlh6/pullmd_gave_claude_code_an_mcp_server_so_it_stops/Read original →

// related

#0001
#0001Agents & tools Google AI Forum2 hours ago
`Antigravity` backlash: editor removed, agent command center takes over
60radar
AntigravityAI coding IDE — Google’s agent-first dev tool
The update shifts from a full IDE into an agent hub. Losing file-first editing weakens it as a daily driver; worth watching before switching workflows.
- The core complaint is concrete: files, editor, terminal, and change tracking no longer feel like one controlled workspace.
- A CLI is framed as a poor replacement for a full IDE. For app debugging, terminal-first flow adds friction.
- Cursor and Windsurf were named as competitors that benefit if Antigravity drops its integrated editor.
- Only 5 posts from 3 participants, so this is an early friction signal, not broad market proof.
Source: discuss.ai.google.dev/t/you-did-not-upgrade-antigravity-Read original →
FIG-0011:1
60radar
FIG-0011:1
#0002
#0002Agents & tools Google AI Forum2 hours ago
`Antigravity CLI` `agy` loses auth session in WSL 2
40radar
Antigravity CLICoding agent CLI — runs terminal workflows via `agy`
WSL 2 runs trigger browser login on every agy execution. Skip WSL use for now if repeated auth breaks your coding loop.
- Repro path is simple: run agy in Ubuntu on WSL 2, complete Windows browser auth, exit, then run it again.
- Expected behavior is a persisted Linux-side session; actual behavior is fresh authentication every run.
- Reported environment is Windows host plus WSL 2 Ubuntu 22.04.5 LTS, so native Linux/macOS behavior is not covered.
Source: discuss.ai.google.dev/t/bug-antigravity-cli-agy-fails-toRead original →
FIG-0021:1
40radar
FIG-0021:1
#0003
#0003Agents & tools Google AI Forum2 hours ago
`Antigravity 2.0` backlash: chat-only agent surface removed core IDE controls
50radar
AntigravityAI coding IDE — Google-backed agent-first dev environment
The complaint is not about model quality; it is about losing the editor, terminal, and file tree. Agent IDEs that remove manual control create churn risk, not lock-in.
- The reported change replaces a classic IDE workflow with a chat-first Manager Surface; that breaks fast component edits, terminal work, and Git control.
- The requested fix is a Legacy IDE toggle: keep agent delegation optional while preserving manual file explorer and terminal access.
- The threat to move to Cursor or Windsurf is the useful signal. Coding agents still need IDE-grade escape hatches for daily product work.
Source: discuss.ai.google.dev/t/antigravity-2-0-forced-update-isRead original →
FIG-0031:1
50radar
FIG-0031:1

PullMD: Self-hosted URL-to-Markdown MCP server for cleaner LLM web reading

// related

`Antigravity` backlash: editor removed, agent command center takes over

`Antigravity CLI` `agy` loses auth session in WSL 2

`Antigravity 2.0` backlash: chat-only agent surface removed core IDE controls