Frequently Asked Questions

BKMRK is bookmark intelligence for developers. It syncs your X/Twitter bookmarks, analyzes each one with Claude AI against your active coding projects, scores relevance, and extracts actionable implementation suggestions.

Instead of a graveyard of forgotten links, you get a prioritized queue of things that actually matter to what you're building.

Two ways: sign in with your X account at bkmrkapp.com, or let OpenClaw create an account for you via the API (see OpenClaw section below).

No. OpenClaw can create your account via the API and start submitting URLs immediately. Connect X later from the dashboard if you want automatic bookmark sync.

Install the BKMRK extension, go to x.com/i/bookmarks, and click SYNC ALL. The extension scrolls through your bookmarks, captures them via X's internal API, and sends them to BKMRK for analysis. Choose a time range (2 weeks, 1 month, 3 months, or all time) to control how far back to sync.

Only Scale users can optionally connect for automatic daily sync via the X API — bookmarks from your phone get picked up without opening a laptop. Pro users import via the Chrome extension from desktop.

Pro: Open your laptop periodically and run a sync from x.com/i/bookmarks — the extension captures everything since your last sync.

Scale: Automatic daily sync via X API picks up phone bookmarks without any action needed.

Fastest way: Install the BKMRK skill from ClawHub and add your API key to ~/.openclaw/openclaw.json.

Then configure your API key (from Settings) in your OpenClaw config:

Once configured, just talk to OpenClaw naturally: "save this URL to bkmrk," "what's in my library about React hooks," or "find my highest-scored bookmarks."

Alternative: OpenClaw can also call the BKMRK API endpoints directly without a skill. Share bkmrkapp.com/agent.json in a conversation and it will figure out the rest.

Yes. Tell OpenClaw you want to use BKMRK and give it your email. It calls /api/agent/onboard, creates your account, and gets an API key back instantly. No OAuth needed. You can connect X later from the dashboard to enable bookmark sync.

Submit any URL — tweets, articles, YouTube videos, GitHub repos, docs. OpenClaw sends it to BKMRK, the enrichment pipeline extracts the content, and Claude analyzes it against your projects.

Query your library — search by keyword, project, priority, or status. Results come back with match scores, implementation suggestions, and Claude Code prompts. Queries are unlimited and free on all plans.

Free users get 5 lifetime trial submissions to try the full pipeline. Paid plans get unlimited submissions (shared with the monthly bookmark cap).

Everything you submit through OpenClaw shows up in your BKMRK dashboard as first-class cards with source badges, scores, and actionable insights.

Any channel OpenClaw supports: Slack, Discord, Telegram, or wherever your agent lives. See an interesting link? Forward it to OpenClaw and it lands in your BKMRK library, analyzed against your projects, by the time you sit down at your desk.

Yes. URLs submitted via agents or the dashboard share the same bookmark cap as synced bookmarks:

Library queries are always free and unlimited.

Every bookmark flows through a simple pipeline: New → Staged → Done.

New — freshly synced or submitted. Claude has analyzed it, but you haven't acted on it yet. This is your inbox.

Staged — you've marked it for action. These are the items you plan to implement, read deeper, or execute with an agent. Think of it as your to-do list.

Done — you've acted on it. Shipped the feature, read the article, applied the pattern. Archive complete.

There's also Trash — for noise you want to discard. Empty the trash to permanently delete items and free up library space.

The sidebar has three filter groups:

Projects — filter by which project a bookmark matched against. Select a project to see only items relevant to that codebase. "All Projects" shows everything.

System Filters — filter by pipeline status. All Items, Unprocessed (new), Staged, Done, and Trash. Counts update in real time as you move items through the pipeline.

Source — filter by where the content came from: X tweets, X Articles, YouTube videos, GitHub repos, web articles, or OpenClaw-submitted URLs. Source filters only appear when you have items of that type.

Filters combine, so you can view something like "Staged YouTube videos for my DIG project."

Critical items are high-priority, high-relevance bookmarks that Claude flagged as especially useful for your projects. These are the ones most worth acting on first.

Yes. Use the + URL button in your dashboard toolbar, or submit URLs through OpenClaw. Any web page, tweet, YouTube video, or GitHub repo works. The enrichment pipeline fetches the full content and Claude analyzes it against your projects, just like a synced bookmark.

Paste any YouTube URL and BKMRK extracts the full video transcript — not just the title or description. This means you can run a 2-hour podcast through Claude and get project-specific takeaways in seconds.

Example: submit a technical podcast episode and BKMRK will tell you which segments are relevant to your projects, extract implementation patterns, and generate actionable suggestions — all without watching the video.

BKMRK extracts the full body text of X Articles (long-form posts on X), blog posts, news articles, and documentation pages. Claude analyzes the entire content — not a preview or snippet — against your projects.

For threads, BKMRK reconstructs the full thread and extracts every URL from every tweet in the chain, running each linked article through the same deep enrichment pipeline. Note: Due to X API limitations, full thread reconstruction works best for threads posted within the last 7 days. For older threads, we fetch the root tweet to give Claude both ends of the conversation.

X tweets — full text, thread reconstruction, all linked URLs enriched

X Articles — full article body via the X API

YouTube — full transcript (auto-generated or manual captions)

Blog posts & articles — full text extraction, uncapped length

GitHub repos — README and repository metadata

Any web page — article text, title, and preview image

All content is sent to Claude uncapped for analysis. Long articles, full podcast transcripts, and detailed technical posts all get the same deep, project-specific treatment.

Each bookmark is enriched (full content extracted, threads unrolled, metadata gathered) then sent to Claude with your project context: name, description, tech stack, and focus areas. Claude scores relevance 0-100%, assigns priority (high/medium/low), categorizes the content, and generates specific implementation suggestions.

Batch sync: Claude Haiku 4.5 for fast, cost-efficient screening during bulk imports.

Re-analyze / deep analysis: Claude Sonnet 4.6 for detailed project-specific analysis with richer suggestions.

A 0-100% relevance score showing how useful a bookmark is to a specific project. Scores above 70% are high relevance (green), 40-70% are medium (amber), and below 40% are low (default). Each project gets its own independent score for the same bookmark.

Click the re-analyze button on any bookmark card in your dashboard. This runs a fresh Sonnet 4.6 analysis against your current projects (useful after updating project descriptions or adding new projects). Re-analyze limits: Free 25/mo, Pro 100/mo, Scale 500/mo.

50 active bookmarks, 1 project, 25 re-analyses/month, 5 agent submissions (lifetime trial), full dashboard access, and the /bkmrk Claude Code skill. No daily auto-sync (one-time import only).

Pro ($9/mo or $79/yr): 300 bookmarks/month, 5 projects, Chrome extension sync, 100 re-analyses/mo.

Scale ($29/mo or $249/yr): 1,000 bookmarks/month, unlimited projects, Chrome extension sync + automatic daily sync via X API, 500 re-analyses/mo.

From the dashboard or pricing section. Payments are handled by Stripe. You can switch between monthly and annual billing or change plans from the billing portal in Settings.

When you hit your bookmark cap, new syncs won't import additional items until you make room (trash + empty trash to hard-delete). Daily sync and monthly import counters reset automatically. URL submissions (from the dashboard or agents) share the same monthly bookmark cap.

In Settings, under the Bookmark SKILL section. Click "Reveal" to see the full key, or "Copy" to clipboard.

Go to Settings → Connected Accounts → Connect GitHub. This lets BKMRK pull project context from your repos for better analysis matching.

Go to Settings → Danger Zone → Delete Account. Type "DELETE" to confirm. This permanently removes your account, all bookmarks, projects, analysis data, and agent usage history. It cannot be undone.