SpiderLook のリモート MCP サーバーとは何ですか?
SpiderLook のリモート MCP サーバーは、https://api.spiderlook.com/mcp にホストされた Model Context Protocol エンドポイントです。AI アシスタントはローカルに何もインストール・実行することなく接続できます。
S
SpiderLook MCP Server
@SpiderLook
基本情報
SpiderLook MCP Server について
SpiderLook collects competitors' social posts, Meta and Google ads and website changes every day and stores the evidence. Its remote MCP server lets Claude, ChatGPT, Gemini, Cursor and other MCP clients answer what changed, why it matters and what to do next. Sign-in uses OAuth 2.1.
接続情報
https://api.spiderlook.com/mcpセットアップ
claude mcp add spiderlook --transport http https://api.spiderlook.com/mcpSpiderLook MCP Server のツール
13Catalog of orgs, projects, and competitors the user can access. Use when project or competitor IDs are unknown. Each competitor includes kind (self|competitor|trendsetter) and is_self.
Batched project counts and named formulas. Not a dashboard narrative. Valid metrics: relative_activity_index, activity_threat, activity_counts, ad_launches, ad_counts, follower_totals, talent_velocity, open_deviations, strategy_shifts, weekly_kpi_series. Prefer counts and formulas; weekly_kpi_series is the uniform competitor×week×channel history surface. strategy_shifts is analyzer text the app already shows. Do not N+1 query_activity to rebuild rankings. window_days is clamped to 1..90 (default 30).
Scraped posts, ads, or pages for one channel, windowed and newest first. Primary evidence tool. Read these items and form your own answer; do not look up Spiderlook's stored channel-strategy writeup. Instagram feed vs reels is variant=feed|reel|all, not a separate channel. window_days is clamped to 1..90 (default 30).
Full stored row, current profile, or snapshot series. kind is post, ad, page, profile, or snapshots. item_id is required for post/ad/page. Profiles and snapshots omit item_id; snapshots honor window_days.
Persisted deviations or strategy versions, newest first, cursor-paged. Prefer kind=deviation (playbook-break events with evidence). kind=strategy_version is a historical dump of Spiderlook's stored LLM writeup — skip it unless the user asks for that history. status applies only to deviations.
Create a tracked company when the user requests adding it to a project. Saves a company record and starts website/profile discovery and favicon retrieval from its public website. Full scanning requires activation. kind=self sets the project's baseline if none exists; competitor is a rival; trendsetter is an inspiration source. Duplicate website hosts produce warnings but still create another record, so retrying can create duplicates. Uses existing account resources; does not purchase subscriptions or credits.
Re-label kind (competitor ↔ trendsetter) or toggle continuous_tracking. Do not use kind=self here — promote baselines via the product self-competitor flow. Active baselines cannot be re-labeled until cleared.
Save a channel and request a crawl when the user asks to add or refresh it. Creates a profile link if absent; reuses matching links. Eligible crawl jobs fetch public channel data and may replace stored analysis. Repeating a call can queue another crawl. Crawls consume available account usage credits and are skipped when scanning is inactive or credits are unavailable; the saved link remains. This tool does not purchase subscriptions or credits. Channels: linkedin, linkedin_person, twitter, instagram, tiktok, youtube, meta_ads, facebook_posts, website, google_ads. facebook_ads and meta alias meta_ads; facebook aliases facebook_posts. Social and ad channels require a matching profile URL; linkedin_person requires linkedin.com/in/..., twitter accepts twitter.com or x.com, and youtube accepts youtube.com or youtu.be. Omit url for website and google_ads; they use the competitor's website.
The team's sales battlecard for one rival. Returns content (null until one is generated), pending suggestions with their ids and evidence, approval state, and whether generation is running.
Generate a battlecard when the user requests creating or refreshing it. Sends stored competitor strategy, the project's baseline strategy if present, and open deviations to the configured AI service. Saves the initial card; later runs preserve existing content and add pending suggestions for review. Requires activated scanning and existing analysed data. Rejects concurrent generation; later retries may create additional suggestions. Uses existing account AI resources and does not purchase subscriptions or credits.
Accept or dismiss one pending battlecard suggestion. Use when the user requests accepting or dismissing a specific suggestion. Accepting may add or replace existing card content and marks the suggestion accepted. Dismissing marks it dismissed without changing card content. suggestion_id comes from get_battlecard pending_suggestions. A suggestion that has already been reviewed is rejected without further changes.
Mark a rival's battlecard as approved by the caller.
Record feedback about a Spiderlook capability unavailable in the listed tools when it prevents completing the user's Spiderlook request. Sends feedback to the developers through PostHog; does not discover or enable additional tools.
概要
What it is
SpiderLook is a competitor-intelligence service. It scans rivals' websites, social posts and Meta and Google ads daily and keeps the evidence, so an AI assistant connected over MCP can report what changed and suggest a response.
What it can do
- Daily scans of competitors' websites, social posts and ads
- Stored evidence behind each change
- Ask what changed, why it matters and what to do next from any MCP client
Connect
Remote endpoint (Streamable HTTP): https://api.spiderlook.com/mcp. Official MCP Registry: com.spiderlook/spiderlook.
Authentication and limits
OAuth 2.1 sign-in; requires a SpiderLook account.
Sources
- Website: https://spiderlook.com/mcp
- Docs: https://spiderlook.com/docs/mcp
よくある質問
SpiderLook の MCP サーバーにはどう接続しますか?
エンドポイント https://api.spiderlook.com/mcp を Claude Code、Cursor、VS Code など任意の MCP 対応クライアントに追加してください。このページのセットアップ手順で各クライアントを一度に設定できます。
SpiderLook の MCP サーバーは認証が必要ですか?
はい。SpiderLook は OAuth を使用します。初回接続時に MCP クライアントがブラウザを開いてサインイン・認可を行い、以降はその認証情報を再利用します。
SpiderLook の MCP サーバーはどのトランスポートを使用しますか?
SpiderLook は streamable-http エンドポイントを公開しており、これはリモート MCP サーバーで広く使われ、主要な MCP クライアントがサポートするトランスポートです。
他のリモート MCP サーバー
3
3GPP Scout
Hosted 3GPP MCP server: cited search across Rel-15 through Rel-20 TS/TR specs, recent TDocs, company proposals and meeting decisions. OAuth 2.1; no local index required.
APIXO AI
Remote MCP for AI image, video, music, speech and custom voice generation. Discover models and input schemas for free; authorize your APIXO account to generate media and check results.
Adgentek Ad Server (AdsMCP)
AdsMCP is the MCP entry point into the Adgentek Ad Server. Call ads_serve with the conversation context and get back a native ad matched to user intent, or no ad when one does not fit. Publishers earn on a CPM basis. Fre
AdWhispr
Research competitors' winning ads, clone them for your brand and launch paused campaigns on Meta, Google and TikTok from chat.
A
AI Pricing Hub
Read-only, source-backed AI model pricing, history, comparisons and cost calculations.
コメント