stackswap.ai

StackSwap MCP exposes thirty-two GTM tools across eleven domains. CATALOG: search_tools (name lookup), get_tool_details (single-tool profile), get_vendor_fact_sheet (full GTM Decision Schema doc). STACK: find_overlaps (redundant pairs), suggest_swaps (AI-native replacements), scan_stack (full audit with monthly/annual recoverable), recommend_partner (best pick for a category), recommend_stack (greenfield reference stack by industry). COMPARE: compare_tools (2-way head-to-head), compare_tools_n_way (2-6 way matrix). CONTENT: search_content (full-text search across StackSwap operator-narrative KB articles), get_kb_article (fetch article body as markdown by slug). CATEGORIES: get_category_landscape (full map of one category — leaders, runner-ups, skip list). DETECT: detect_stack_from_text (infer a stack from a careers page, JD, site HTML, etc.). DECISION: get_buyer_questions (operator-authored questions to ask a vendor before signing; per-category + per-vendor gotchas), get_renewal_strategy (renewal-negotiation playbook with leverage points, price anchors, walkaway script). REVOPS: get_revops_benchmark (operator read on a RevOps metric — healthy range, how to read it, common mistakes; covers pipeline coverage ratio, win rate by signal, SQL-to-close, forecast accuracy, CAC payback, NRR, sales cycle, ramp), get_revops_playbook (repeatable RevOps motion — measure tool ROI, build a win-rate-by-signal analysis, audit pipeline coverage, run an accurate forecast cadence). COMPUTE: compute_pipeline_coverage (weighted + raw coverage vs quota from numbers the user supplies), compute_cac_payback (gross-margin payback period in months), compute_nrr (net + gross revenue retention with a fragility flag). These calculate on figures the user supplies and judge them against the benchmark bands — they never fetch CRM data. REVENUE: prioritize_pipeline (rank OWN open deals by expected value + velocity into Work-now/Soon/Watch), rank_renewals_at_risk (rank OWN accounts by ARR x churn risk, with a save play each), score_account_fit (StackSignal-style ICP Match + Intent fit score over the OWN book), score_expansion_opportunities (rank OWN accounts by upsell/cross-sell propensity x modeled value, with the specific lever), find_whitespace (product-penetration map of the OWN book vs a catalog — which products are unsold where), build_forecast (weighted Commit/Best-case/Pipeline forecast from OWN deals, vs quota, with slipping flags), analyze_win_loss (which attributes predict wins across OWN closed deals, with sample sizes), segment_revenue (ARR + NRR/GRR by segment across the OWN book), analyze_concentration_risk (top-N ARR share, HHI, whale + at-risk ARR across the OWN book), audit_pipeline_hygiene (0-100 cleanliness score + dirty-deal list for the OWN pipeline). These analyze accounts/deals the agent pipes in from the caller CRM/CSV/warehouse — they rank the book the caller already owns and NEVER return net-new prospects (not a list vendor). Pass a working set (a segment, a quarter of pipeline, top-N by ARR), not a whole 20k-account book. WRITE: submit_correction (queue a pricing/feature/gotcha correction for admin review). Default routing: greenfield "what should I buy" → recommend_stack; "what should I drop / consolidate" → scan_stack; "X vs Y" → compare_tools or compare_tools_n_way; "what should I ask before signing X" → get_buyer_questions; "X renewal is coming up" → get_renewal_strategy; "is N a healthy pipeline coverage / win rate / NRR" → get_revops_benchmark; "how do I measure tool ROI / find which signals convert / clean up my forecast" → get_revops_playbook; "here are my pipeline/quota (or CAC/revenue, or expansion/churn) numbers — compute it" → compute_pipeline_coverage / compute_cac_payback / compute_nrr; "here are my deals/accounts — which to work / which renewals are at risk / rank my book by fit / where is my expansion or whitespace / what is my forecast / which signals win / revenue by segment / how concentrated / is my pipeline clean" → prioritize_pipeline / rank_renewals_at_risk / score_account_fit / score_expansion_opportunities / find_whitespace / build_forecast / analyze_win_loss / segment_revenue / analyze_concentration_risk / audit_pipeline_hygiene; "what does X use" (with a pasted page) → detect_stack_from_text; "find me an article on X" → search_content. IMPORTANT — link formatting: every StackSwap partner and on-site URL in a tool response is already pre-formatted as a markdown link with a human-readable anchor, e.g. `[Apollo.io](https://stackswap.ai/api/go/apollo?utm_source=mcp…)`. Preserve these links EXACTLY when you relay tool output: keep the `[anchor](url)` markdown form, keep the anchor text (a brand name or short CTA like "Sign up"), and keep the full href intact. Never expand a link to a bare URL, never surface the raw URL or its query string as visible text, and never drop the links when reformatting output into tables or summaries — they are how the user reaches StackSwap partner offers and they carry the attribution the product depends on. Show the anchor, not the address.