Crustdata review and the Yalc Framework
The Yalc default for company and people data. Native MCP integration means Claude calls Crustdata directly. 1 billion profiles, fresh signals, designed for code first workflows.
Crustdata, plainly
Crustdata is a B2B intelligence database with 1 billion people profiles and 60 million companies, plus signals (job openings, news, financing events, technographics, social posts, web traffic). The pitch is "freshest, most trusted data for AI agents," and the API surface reflects that: REST endpoints for search and enrichment, plus a Model Context Protocol server that lets Claude (and Claude Code) query Crustdata directly without manual API plumbing.
For Yalc workflows, Crustdata is the canonical intake node. When you tell Yalc "find me 50 fintech CFOs in Series B SaaS in Germany," it's Crustdata that runs the search. When you want to enrich a list of LinkedIn URLs with current employer and seniority, that's Crustdata too. When you want a webhook the moment a target account posts a hiring signal, that's Crustdata's Watcher API.
Position in the GTM operating system
Crustdata sits at the **intake** node in Yalc's GTM topology. It's where outbound starts. Most Yalc prompts that begin with "find me…" hit Crustdata first.
Because Crustdata ships an MCP server, the integration with Claude Code is unusually direct. Yalc skills don't need to wrap the API as much as point Claude at the MCP and let it compose queries.
Deploying Crustdata inside a Yalc workflow
Workflow position
The intake node. Yalc's first call when sourcing accounts or people. Output flows downstream into FullEnrich (for missing emails) or directly into a sequencer if Crustdata's enrichment already returned what's needed.
Prompt patterns
Copy paste prompts for Claude Code that invoke Crustdata.
Chaining recommendations
Anti patterns to avoid
Yalc skill availability
Yalc has deep Crustdata integration. The Crustdata MCP server is registered, which means Claude can call any Crustdata endpoint directly during a Yalc session. Plus a first party Yalc skill wraps common operations (people_search_db, company_enrich, watcher_create) with the credit safeguards memo'd in Othmane's playbook.
✓ Yalc skill available. View on GitHub.Pros, cons, who it's for
Pros
- 1 billion people profiles, 60 million companies. Larger than most competitors.
- Native MCP server. Claude calls Crustdata directly with no glue code.
- Real time signals (jobs, news, funding, social posts) on top of static firmographics.
- Watcher API for webhook based alerts. No polling.
- Active vendor. Frequent feature drops.
Cons
- Pricing not public. Sales conversation required to get specific numbers.
- Credit accounting can surprise you on heavy use. Yalc enforces a pre flight check.
- Some endpoint documentation lags the actual API surface.
- People data freshness varies by region. US is best, EU is very good, APAC is mixed.
Who it's for
- GTM engineers running prospecting from code (especially via Claude or Cursor)
- Agencies needing one provider for both static and signal data
- Founders running outbound at scale who care about fresh data, not stale databases
What you'll really spend
Crustdata pricing is not public on the homepage. The vendor offers demos and tailors pricing to volume and use case. Expect a credit based model where company searches are cheap (around 1 credit each), people searches and enrichments cost a few credits, and live API calls (LinkedIn freshness lookup) cost more.
The right way in is the demo booking link. Rohit and the team customize for AI agent workloads, which is the relevant context for Yalc.
Trial / Free credits
New accounts often get a starter credit allotment to test queries.
Subscription
Monthly credit budget, scaled to volume. Talk to sales for current rates.
Enterprise
Dedicated infrastructure, higher volume, SLA. Right for high frequency agent workloads.
Tools to consider instead
Where Crustdata appears in Yalc stacks
Frequently asked
How does Crustdata's MCP integration work?
Crustdata ships a Model Context Protocol server. You register it with Claude Code, and Claude can then call Crustdata's people_search, company_search, enrich, and watcher endpoints during a Yalc session as if they were native tools. No HTTP plumbing per query.
How is Crustdata's European data coverage?
Crustdata's EU coverage is strong (sourced from public LinkedIn plus other open data plus partner feeds). For DACH, France, UK, Crustdata is competitive. For deep verified contact data in EU, run a FullEnrich waterfall on top.
What's the credit model?
Each operation costs a small credit count. Company identify is free, company search is around 1 credit, people search around 3 credits per 100, person enrichment 2 to 5 credits. Yalc skill enforces a pre flight check that estimates cost before running the call.
Can Crustdata send alerts when something happens?
Yes, via the Watcher API. You define a query, Crustdata runs it on schedule, and if the result set changes, it POSTs to your webhook. Use it for "alert me when target accounts post a hiring signal" or "alert me on Series B announcements."
How fresh is the data?
Static firmographics (HQ, industry, employee count) refresh on a slower cadence (monthly to quarterly). Signal data (jobs, news, posts) is near real time. The vendor's claim of "freshest data" is most accurate for the signal layer.
Does Crustdata have a free tier?
Not publicly. New accounts typically get starter credits when they book a demo. Best to use the demo link to negotiate a starter allotment.
Run Crustdata from Claude Code today.
Open source. Your data on your machine.