ApparelHub vs DataDive
A side-by-side look at the ApparelHub and DataDive MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose ApparelHub for e-commerce work as a local process; choose DataDive for e-commerce as a local process. You can also run both side by side in one mcpServers block.
Which MCP server should you choose?
Choose ApparelHub when
- You need dedicated capabilities in the E-Commerce domain.
- You prefer a local stdio subprocess transport.
- Its auth model fits you: OAuth 2.0.
- You have the keys it needs:
APPARELHUB_API_KEY. - You want tools like
check_setup_readiness, list_connectable_providers, connect_fulfillment_provider.
Choose DataDive when
- You need dedicated capabilities in the E-Commerce domain.
- You prefer a local stdio subprocess transport.
- You have the keys it needs:
DATADIVE_API_KEY. - You want tools like
list_niches, get_niche_keywords, list_keyword_exclusions.
Feature & specification comparison
Rows where the two servers differ are marked.
| Spec | ||
|---|---|---|
| Category | E-Commerce | E-Commerce |
| Quality signal | 63/100 · Good | 63/100 · Good |
| Install command | Confirmed | Confirmed |
| Transport | Local subprocess (stdio) | Local subprocess (stdio) |
| Auth | OAuth 2.0 | Not declared |
| Pricing | BYOK (pay the provider) | Not declared |
| Required env vars | APPARELHUB_API_KEY | DATADIVE_API_KEY |
| Verified owner | Community listing | Community listing |
| GitHub stars | 0 | 1 |
| Directory activity | 1 views 0 installs 0 upvotes | 1 views 0 installs 0 upvotes |
| Tools | 124 | 43 |
| Full listing | View ApparelHub | View DataDive |
Tools & capabilities
ApparelHub tools 124
check_setup_readinessWhat this account already has, what it still needs, and the single next action to take. Returns ready_to_design / ready_to_fulfill / ready_to_sell, a per-store breakdown, and an ordered next_steps list. Start here for any first-time setup, and call it again after each connection to confirm the state actually changed. Read-only, makes no provider calls, and is safe to poll. [#01f20c]list_connectable_providersFulfillment providers and sales channels this account may connect, each marked with how it connects: connect_mode "in_chat" means you can complete it here by asking for a credential, "browser" means you must dispatch an authorization link with start_channel_connect and poll. Also returns where the merchant generates the credential, when there is one. Use this before asking a user for anything, so you ask for the right thing. [#21af2f]connect_fulfillment_providerConnect an API-token fulfillment provider (Printify, Gelato) to a store, entirely in chat. Validates the token first, so a bad token fails before anything is stored. If the token maps to more than one shop the result asks you to pick one and lists them — call again with shop_id set. For Printful use start_channel_connect instead: it needs a browser. Never repeat the token back to the user. [#98598c]connect_sales_channelConnect an API-key sales channel (WooCommerce, Wix) to a store, entirely in chat. For Shopify and TikTok Shop use start_channel_connect instead: they need a browser. Credentials are write-only and are never returned. [#c9093d]start_channel_connectBegin a browser-based connection (Printful, Shopify, TikTok Shop, Fourthwall). Shopify additionally requires shop_url, the merchant myshopify domain — ask for it before calling. Returns an authorization URL to give the user. THE CONNECTION IS NOT FINISHED WHEN THIS RETURNS. You must keep polling check_connection_status (passing the same provider_uuid) until it reports connected, then tell the user. The browser tab where they authorize is NOT this conversation and cannot report back to you, so polling is the only way you or they will learn it worked. Poll every few seconds, up to about two minutes, and if it has not landed by then ask whether they finished authorizing rather than giving up silently. If they need to create an upstream account first, let them, then call this again for a fresh link. [#6d4006]check_connection_statusPoll whether a dispatched connection has completed. Call this repeatedly after start_channel_connect while the user authorizes in their browser, and announce the result when it lands: they cannot see this conversation from the tab they authorized in, so if you do not tell them, nobody does. Read-only, makes no provider call, and is safe to poll every few seconds. connected true means say so and continue setup. connected false means keep waiting. needs_reconnect means retrying will never work and you must dispatch a fresh link with start_channel_connect. [#807361]list_my_workspacesList the workspaces this account can act in, each with its uuid. Agency / multi-brand accounts have more than one (e.g. a workspace per client); a single account just has Default. The store / product / order / design tools operate on the Default workspace unless you pass workspace=<uuid>. Use this FIRST to resolve a workspace by name (e.g. a client's name) to the uuid those tools need. Read-only. [#358d40]list_my_storesList the merchant's ApparelHub stores, each with its fulfillment providers (Printful/Printify) and connected sales channels (Shopify/WooCommerce/Wix). Read-only. [#739377]list_my_designsList the merchant's generated design images (newest first). Read-only. Use these design_uuids with the design/product tools. Pass on_products=false to find orphan designs (designs not used by any live product), the supported way to audit a workspace for unused designs before archiving them. Pass archived=true to list already-archived designs. A design with no full_url carries processing_status: "pending"/"processing" means it is still being made and is worth polling, while "failed" means it gave up and processing_error says why. Branch on processing_error_code rather than matching the message text, and do not retry a design whose failure is a content block — it will fail the same way every time. [#58124e]get_design_collectionList the OTHER designs that came out of the same generation as this one. When generate_image is called with concepts > 1, one generation produces a themed set; every design is saved separately and this is how you find its siblings. The design you ask about is never included, and the list is oldest first. An ordinary design returns an empty list, which is the honest answer, not an error (an inaccessible or unknown uuid also reads as empty, so an empty list does not prove the design exists). Read-only and not metered. Each entry carries title and concept_description: the product name and retail description the model wrote for that concept. Use them to build products without asking the merchant to retype copy the model already produced; either may be null and must never be invented. background_state says whether the design can be lifted off its background: transparent is ready to place, opaque_box can be keyed automatically (process_transparency), opaque_varied cannot be separated and will print WITH its background. Asked for a set, the model often answers with pictures of finished garments rather than flat art, and building a product from one prints a garment onto a garment, so do NOT build a product from an opaque_varied design without telling the user it will print with its background. Do not describe it as "a photograph" either: legitimately textured art reports opaque_varied too. [#f2d543]list_my_productsList the merchant's products with their fulfillment and sales-channel sync status. Each channel entry also carries `health` — what the channel last said about the listing. A channel can remove or deactivate a listing at any time, so check `health`, not just `sync_status`: a product can read 'Synced' historically and still be gone. Pass store_uuid to scope to one store; omit for all products. Read-only. [#06afd6]list_my_ordersList the merchant's recent orders across channels. An order carrying `needs_attention` is one the merchant must act on (the fulfillment provider failed or cancelled it, the buyer cancelled after production started, or it was paid but never sent to production): always tell the user about those, with the `message`. Read-only. [#20d0e7]- +112 more on the listing page
DataDive tools 43
list_nichesUse this first when the user asks about their niches, or to find a `nicheId` for use with `get_niche_keywords`, `get_niche_competitors`, or `get_ranking_juice`. Retrieves a paginated list of Niches, newest dive first. Each Niche represents a market segment or product category being tracked. Returns nicheId, heroKeyword, nicheLabel, marketplace (com/uk/de/...), and latestResearchDate per niche, plus pagination metadata (currentPage, pageSize, total, lastPage, hasNext, hasPrev). Narrow with `searchText` (label/keyword) or `searchAsin` (a competitor ASIN) instead of scanning pages; a page holds at most 50 niches, so the full list of a large account needs several calls.get_niche_keywordsUse this when the user asks about keywords, search terms, or search volume for a niche. Retrieves the master keyword list for the specified Niche — the relevant search terms monitored for ranking and performance metrics. For each keyword, returns search volume, relevancy (numeric score or "Outlier"), and competitor ASIN organic ranks (asinRanks: { ASIN -> rank | null }), plus `origin`: `mkl` (found by the dive), `manual` (added by hand with `add_niche_keywords`), `outlier` / `residue` (moved in from those lists). Only `manual`, `outlier` and `residue` keywords can be taken out with `remove_niche_keywords`. Keywords hidden by the niche's keyword exclusions are not listed; see `list_keyword_exclusions`. Also returns the latestResearchDate.list_keyword_exclusionsUse this to see which keywords are hidden from a niche's Master Keyword List, e.g. when a keyword the user expects is missing from `get_niche_keywords`. Returns the same lists the DataDive web app shows, oldest first: `exact` — each entry hides the one keyword that matches it exactly, with that keyword's `searchVolume` (0 when the niche has no such keyword); `phrase` — each entry hides every keyword that contains the phrase, with `excludedKeywordsCount` and the total `searchVolume` of the keywords it hides (not counting ones already excluded exactly). Exclusions hide keywords wherever they come from: the dive, the Outliers and Residue lists, and hand-added ones. Change them with `add_keyword_exclusions` and `remove_keyword_exclusions`.get_niche_rootsUse this when the user wants to find the highest-impact words across a niche's keywords, asks about 'keyword roots', or wants to prioritize terms for a listing. Retrieves the keyword lexical roots for the Niche — individual words or word-combinations extracted from the Master Keyword List (e.g. 'bluetooth headphones' yields roots 'bluetooth', 'headphones', 'bluetooth headphones'). Returns `roots` and `normalizedRoots` tables, each item with root, frequency (how many keywords contain it), broadSearchVolume (summed search volume of those keywords), and broadSearchVolumeRatio (0-1, relative to the top root), plus the per-keyword breakdowns and latestResearchDate.get_niche_competitorsUse this for competitor ASINs, product titles, BSR, sales/revenue benchmarks, or niche opportunity scoring. Retrieves the list of Competitors within the specified Niche along with Niche statistics. For the Niche: keyword statistics, opportunity evaluation, benchmark median values, and overall competitor strength assessment. For each Competitor: ASIN, product title, BSR, category and full category tree, sales, revenue, ratings, reviews, price, number of variations, image URL, total Ranking Juice, and ranking data. BSR supports ranking Competitors by sales position within mentioned category. For the Ranking Juice breakdown by title, bullets and description, use `get_ranking_juice`.get_ranking_juiceUse this when the user asks 'how do I rank higher', 'what's my ranking juice', or wants listing-optimization guidance. Retrieves the Ranking Juices for each Competitor within the specified Niche. Ranking Juice is DataDive's proprietary metric identifying the key factors driving search-ranking success for each Competitor's listing. Returns currentListing and optimizedListing breakdowns (rankingJuice + per-property contributions: title, bullets, description) plus per-competitor listing scores and the latestResearchDate.list_rank_radarsUse this to find a `rankRadarId` before calling `get_rank_radar_data`. Filter by `nicheId` if the user has already identified a niche. Retrieves a paginated list of Rank Radars — keyword-rank trackers monitoring organic and sponsored positions for specific ASINs over time. Each item includes id, status (ACTIVE, PAUSED or ARCHIVED), asin, marketplace, keywordCount, title, imageUrl, and summary metrics: top10KW, top10SV, top50KW, top50SV. Supports pagination metadata (currentPage, pageSize, total, lastPage, hasNext, hasPrev); a page holds at most 50 Rank Radars, so a large account needs several calls.get_rank_radar_dataUse this to analyze keyword ranking trends over time. Requires startDate and endDate (yyyy-mm-dd), at most 90 days apart. Retrieves historical keyword ranking data for the specified Rank Radar within the date range. Without `currentPage` and `pageSize` it returns every active keyword in one result. To read one page instead — faster for large Rank Radars — pass `currentPage` and/or `pageSize` (max 100) and continue with `currentPage` + 1 while `hasNext` is true. The response carries currentPage, pageSize, total, lastPage, hasNext and hasPrev; `total` is the Rank Radar's active keyword count. Paused keywords are not included. Each keyword has id, keyword, searchVolume, relevancy, ranks (per-day { date, organicRank, sponsoredRank, impressionRank }), and `highlights` — the user tags on this keyword. Each `highlights` entry is the same object `list_rank_radar_tags` returns (id, title, color, description, startAt, endAt, type, krtId, krtKeywordId); its `id` is the `tagId` the tag-management tools take. A rank of 101 means the product was not in the top 100 results that day. For PPC metrics use `get_rank_radar_ppc_data`; for Search Query Performance use `get_rank_radar_sqp_data`. Use after `list_rank_radars` to discover a `rankRadarId`.get_rank_radar_sqp_dataUse this when the user asks how shoppers search, click, add to cart and buy for the keywords of a Rank Radar — Amazon's Search Query Performance (SQP) data, from Brand Analytics. Requires startDate and endDate (yyyy-mm-dd), at most 90 days apart; 30 days or less keeps it fast. Each keyword's metrics are aggregated over the range: searchQueryVolume and searchQueryScore; impressions, clicks, cart adds and purchases, each as the total across all sellers (`*TotalCount`), this ASIN family's count (`*AsinCount`) and its share (`*AsinShare`); click, cart-add and purchase rates; and ctr/cvr for the market (`*Total`) and for this ASIN family (`*Asin`). `numberOfDaysWithData` says how many days of the range had SQP data; metrics are null when there is none, which is normal for low-volume keywords, recent dates (Amazon publishes SQP with a delay) and sellers without Brand Analytics. Results are paged by keyword: `data` holds one page, and the response carries currentPage, pageSize, total, lastPage, hasNext and hasPrev. `total` is the Rank Radar's active keyword count; to read every keyword, call again with `currentPage` + 1 while `hasNext` is true, using `pageSize` 100 for large Rank Radars. Paused keywords are not included. Each call counts toward API usage and is rate limited (about 60 requests/minute). Use after `list_rank_radars` to discover a `rankRadarId`.get_rank_radar_ppc_dataUse this when the user asks about advertising (Sponsored Products) performance for the keywords of a Rank Radar. Requires startDate and endDate (yyyy-mm-dd), at most 90 days apart; 30 days or less keeps it fast. Each keyword's metrics are aggregated over the range: sponsoredRank (median; 101 means not in the top 100 sponsored results), impressionRank and impressionRankShare; how many exact, phrase, broad and auto campaigns target it; organicSales, ppcSales, ppcSpend, costPerClicks, clickThroughRate, conversionRate, totalClicks, totalImpressions, totalOrders and acos. Metrics are null or 0 when the seller runs no ads on a keyword. Set `includeCampaigns` for the per-campaign breakdown. Results are paged by keyword: `data` holds one page, and the response carries currentPage, pageSize, total, lastPage, hasNext and hasPrev. `total` is the Rank Radar's active keyword count; to read every keyword, call again with `currentPage` + 1 while `hasNext` is true, using `pageSize` 100 for large Rank Radars. Paused keywords are not included. Each call counts toward API usage and is rate limited (about 60 requests/minute). For the seller's campaigns as a whole, use `list_ppc_campaigns`. Use after `list_rank_radars` to discover a `rankRadarId`.list_rank_radar_tagsUse this to find a `tagId` before calling `update_rank_radar_tag`, `delete_rank_radar_tag`, `add_rank_radar_tag_search_terms` or `remove_rank_radar_tag_search_terms`. Returns the user tags ("highlights") on the specified Rank Radar, ordered by title (case-insensitive, ascending). Each row carries `id`, `title`, `color`, `description`, `startAt`, `endAt`, `createdAt`, and `keywordCount` — the number of Rank Radar keywords this tag is applied to. The system "paused" tag (the one the quota pauser writes with a reserved colour) is not listed and cannot be addressed from these tools. Supports pagination metadata (currentPage, pageSize, total, lastPage, hasNext, hasPrev); a page holds at most 50 tags.create_rank_radarUse this to start tracking organic and sponsored keyword rankings for an ASIN within a niche. ⚠️ Spends Search Term tokens (cost scales with `numberOfKeywords`) and cannot be undone — set `confirm: true` only after the user approves the cost. Not safe to retry: each call spends tokens again and creates a separate Rank Radar — if a call errors or times out, check `list_rank_radars` instead of re-calling. The first call previews the creation and may return `warnings` alongside the cost note — for example that this product family is already tracked by another Rank Radar. Show every warning `message` to the user verbatim, before asking them to confirm; do not summarise it or decide on their behalf. A warning never blocks creation. Returns the new `rankRadarId`; read its data later with `get_rank_radar_data`. Requires `asin`, `numberOfKeywords` (min 1), and a `nicheId` from `list_niches`.- +31 more on the listing page
Ready-to-paste client configs
Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
ApparelHub
{
"mcpServers": {
"apparelhub": {
"command": "npx",
"args": [
"-y",
"@apparelhub/mcp-server"
],
"env": {
"APPARELHUB_API_KEY": "YOUR_APPARELHUB_API_KEY_HERE"
}
}
}
}
DataDive
{
"mcpServers": {
"datadive": {
"command": "npx",
"args": [
"-y",
"@datadive-tools/mcp"
],
"env": {
"DATADIVE_API_KEY": "YOUR_DATADIVE_API_KEY_HERE"
}
}
}
}
Frequently asked questions
ApparelHub is categorized under E-Commerce and uses a local stdio subprocess. DataDive belongs to E-Commerce and uses a local stdio subprocess. Pick ApparelHub for e-commerce work and DataDive for e-commerce.
Copy the mcpServers block from the config section on this page into your client's config (for example claude_desktop_config.json or ~/.cursor/mcp.json), then fully restart the client.
ApparelHub is bring-your-own-key: usage is billed by the API provider you supply a key for. Auth: oauth 2.0. DataDive has no declared pricing — check the project's own site. If a server needs environment variables such as API keys, define them under the env key in your client's MCP config.
Yes. MCP clients support multiple servers: add apparelhub and datadive as separate keys in the same mcpServers object, and your assistant will call tools from either one.
On AllMCPs, ApparelHub has a quality signal of 63/100 (Good) with 1 views, 0 installs and 0 GitHub stars. DataDive has 63/100 (Good) with 1 views, 0 installs and 1 GitHub stars.
