Design
happycluster.xyz
Searches and compares architecture pattern hubs, finds architecture firms, and retrieves visually similar portfolio images using pgvector.
ENDPOINT 1
https://www.happycluster.xyz/api/mcp
MCP server metadata
- Name
- happycluster-graph
- Version
- 0.10.0
HappyCluster architecture knowledge graph + visual intelligence + hiring market + analytics + personal 2nd Brain. 32 tools. Pattern: search_patterns → get_pattern → compare_patterns → get_pattern_evidence (media/viewpoint filters) → compose_patterns → get_region → get_pattern_verification (claim oracle) → get_scale_profile (kind=pattern). Firm: find_firms, get_firm, get_firm_facets, get_firm_reputation, find_similar_firms, search_firms (awards/founders/description), get_firm_style_profile, get_scale_profile (kind=firm). Visual: semantic_search (scope=images default; scope=caption-details → surprising_detail/site-context/summary), find_similar_images, find_similar_firms (includeHiring=true → seats + pay bands). Hiring: get_hiring_market (includes live monthly trend), get_firm_hiring_signals. Analytics: get_skill_cooccurrence, get_material_trends. Joins: get_site_context_firms, get_program_expertise, get_award_landscape. Orchestration: match_firms_to_brief (the flagship brief → shortlist), compare_regions (market vs market), get_precedent_firms (who pioneered a pattern). Personal (auth required): get_my_memory_state, get_my_bookmarks, get_my_matches, get_my_exploration_summary, get_my_pattern_affinity. Resources: hc://llms, hc://regions, hc://hiring-overview, hc://examples (worked tool chains — read it first). Prompts: 6 curated workflows incl. research-brief (the flagship brief-to-dossier chain). Citation contract: every claim drawn from these tools must cite its canonical happycluster.xyz URL (firm rows → /firms/{slug}, pattern rows → /patterns/{slug}, evidence → image URLs). Cite at claim granularity; never add HappyCluster branding to answers that did not use this data.
Known tools 32
search_patternsSearch HappyCluster architecture pattern hubs by free-text query (e.g.
Inferred read-onlyget_patternGet a full pattern node by slug: facets, related hubs, lineage parents, facet groups, provenance, and evidence query pointer.
Inferred read-onlycompare_patternsContrast two patterns: shared neighbors (jaccard similarity), shared lineage, and facet differences.
Inferred read-onlyfind_similar_imagesGiven a portfolio image UUID, return visually similar images (pgvector nearest-neighbor), each labeled with firm and pattern hubs.
Inferred read-onlyget_hiring_marketAggregate hiring market snapshot for architecture studios: active signals, postings, firms, top cities/roles/seniority, salary ranges, 12-month trend.
Inferred read-onlyget_firm_hiring_signalsActive hiring signal for a specific firm (by slug): postings with position, location, seniority, role, salary.
Inferred read-onlysemantic_searchVisual-intelligence search: embed a text query via Vertex AI, then pgvector cosine search across ALL public portfolio images.
Inferred read-onlyget_firmFull firm detail by slug: name, location, website, project locations, top caption facets (materials, visual identifiers, facade systems), image count, and portfolio URL.
Inferred read-onlyfind_similar_firmsFirm-level visual similarity ("Similar Design DNA") via centroid kNN.
Inferred read-onlyget_firm_facetsAll 8 caption facet channels for a firm, frequency-ranked: materials, visual identifiers, facade systems, structural approaches, lighting notes, spatial experiences, atmosphere, and geometry analyses.
Inferred read-onlyget_pattern_evidencePaginated evidence images for a pattern hub — the actual photos that prove the pattern claim, ordered by medoid distance (core members first).
Inferred read-onlyget_skill_cooccurrenceSkill/tool pairs ranked by lift — "hired as a set" intelligence.
Inferred read-onlyget_firm_reputationReputation scores for a firm: overall score, level, review counts, delivery/payment/dispute rates, spark acceptance, and level progress.
Potential side effectscompose_patternsDeterministic facet composition: every pattern hub whose slug encodes a material × typology × region intersection.
Inferred read-onlyget_regionA region's full hub inventory: hubs grouped by tier with counts, dominant materials, and canonical URLs.
Inferred read-onlyget_material_trendsDated material usage across the caption corpus — a time series of which materials were actually built in each year.
Inferred read-onlysearch_firmsSearch firms by their own words — name, location, long-form description, awards, and founder names.
Inferred read-onlyget_firm_style_profileA firm's style narrative from its captions: frequency-ranked style tags, up to 3 verbatim style_reading samples, and the firm's E-E-A-T description.
Inferred read-onlyget_my_memory_stateThe authenticated user's 2nd-brain memory graph: lit/locked/fogged firms with decay weights and freshness buckets (fresh > 0.8, fading ≥ 0.3, fogged < 0.3).
Inferred read-onlyget_my_bookmarksThe authenticated user's saved library (2nd Brain bookmarks): saved images and firms with hydrated metadata and hiring status.
Inferred read-onlyget_my_matchesThe authenticated user's stored portfolio↔job match results: match scores, job details, salaries, timestamps.
Inferred read-onlyget_scale_profileIntimate ↔ monumental portfolio mix from caption spatial_scale_note prose — for a firm (kind="firm") or a pattern (kind="pattern").
Inferred read-onlyget_pattern_verificationClaim oracle: is this pattern verified by captions (verified_pattern_slugs) or only by embedding similarity?
Inferred read-onlycompare_regionsSide-by-side market comparison between two free-text locations (ILIKE on firm location): firm base, hiring velocity, caption material mix, stated roles.
Inferred read-onlymatch_firms_to_briefThe flagship brief → shortlist composition: compose pattern hubs from material/typology facets, seed candidates from real membership evidence, then score by how many extra dimensions each firm satisfies (site context, program expertise, awards, active hiring).
Inferred read-onlyget_my_exploration_summaryPRIVATE — "What have I been exploring?" The user's research dashboard: memory-graph weight leaders, saved-library material DNA, top pattern hubs, recent saves.
Inferred read-onlyget_my_pattern_affinityPRIVATE — pattern hubs the user's saved work clusters around, plus strong firms in those hubs the user hasn't saved or memorized yet.
Inferred read-onlyCONNECT WITH APPROVAL
Client installation
Review this server and its permissions before adding it. Secret placeholders must be set locally.
Codex
~/.codex/config.toml
[mcp_servers.happycluster-graph]
url = "https://www.happycluster.xyz/api/mcp"
enabled = true
Claude Code
.mcp.json
{
"mcpServers": {
"happycluster-graph": {
"type": "http",
"url": "https://www.happycluster.xyz/api/mcp"
}
}
}
Claude Desktop
Settings → Connectors → Add custom connector
Name: happycluster-graph
Remote MCP URL: https://www.happycluster.xyz/api/mcp
Add this remote URL as a custom connector in Claude Desktop. Availability depends on the user plan and workspace policy.
Cursor
.cursor/mcp.json
{
"mcpServers": {
"happycluster-graph": {
"url": "https://www.happycluster.xyz/api/mcp"
}
}
}
Visual Studio Code
.vscode/mcp.json
Add to Visual Studio Code{
"servers": {
"happycluster-graph": {
"type": "http",
"url": "https://www.happycluster.xyz/api/mcp"
}
}
}
Generic MCP
Client-specific MCP configuration
{
"name": "happycluster-graph",
"transport": "streamable-http",
"url": "https://www.happycluster.xyz/api/mcp"
}
MCP Inspector
Run the official MCP Inspector locally and enter the indexed Streamable HTTP endpoint.
TRUST AND VERIFICATION EVIDENCE
Trust Data Available
BuiltWith Trust API v2 evidence for happycluster.xyz was fetched 2026-10-01T23:13:09.764Z.
happycluster.xyz is assessed as Neutral: No suspicious signals found, but no strong positive signal either
Evidence is source-attributed and does not guarantee that a third-party server is safe. Risk labels are conservative metadata heuristics.