You're about to bet on a market, a deal, or your career.
A bad call costs $50K–$500K+. This costs $0–$500. And it lands in 15 min – 4 h depending on the engagement, not 6 weeks.
$500 protected against $50K–$500K·1/200 the time at 1/500 the cost of a tier-1 engagement
Early access · MCP server · works with any MCP-compatible AI client

You have the decision.
You don't have the analysis.

A 6-week MBB engagement, compressed to 15 minutes to 4 hours depending on scope. 27 specialised agents and 138 atomic tools run inside your AI client of choice — Claude, ChatGPT, Cursor, Codex, Gemini, anything that speaks MCP. 5-lens framework. 10 hypotheses tested. Every number source-cited — verdict ready before your next meeting.

Install & run my first engagement →
Bring your own AI:
Claude ChatGPT Cursor Codex VS Code Copilot Gemini Code Assist Cline Windsurf + any MCP client
You decide on vibes because the alternatives are worse: → MBB firm: $50K–$300K, 6 weeks, lands after the deadline. → A bare LLM prompt: confident slop, zero hypotheses, zero citations. → A consultant friend: one opinion, no portfolio view, no defence at the board.

This isn't another LLM wrapper.

This is consulting — codified.

Use this when you're:
→ closing an investment round → entering a new market → validating an M&A target → pivoting your career
$ claude mcp add co-founder-engine --url "https://mcp.xataco.com/mcp"
one command · OAuth login
// the problem

Strategy is broken at the founder layer

Tier-1 consulting is unreachable. Generic AI is unrigorous. You decide on vibes.

Tier-1 firms are unaffordable

$50,000+ per engagement, 6+ weeks of slide decks. Out of reach for founders, mid-cap operators, and ambitious individuals.

Generic AI gives you slop

One-shot prompts produce one-shot answers. No methodology, no hypotheses, no source verification, no structured analysis.

Document generators ≠ engagements

Existing AI strategy tools spit out PDFs. You don't get a thinking process, a market map, or a portfolio matrix you can defend.

// how it works

Connect once, run engagements in 30 seconds

01

Connect

Add the MCP server to your AI client of choice — Claude, ChatGPT, Cursor, Codex, anything that speaks MCP. OAuth or bearer token. Zero config.

02

Brief

One sentence. Company name, deal target, or your background. The agents ask follow-ups when needed.

03

Engagement runs

27 specialised agents and 138 atomic tools, 5-lens framework, 10 testable hypotheses, source-verified data points.

04

Receive artefacts

Markdown reports, Growth-Share matrix, GTM playbook, validation report — all delivered to a run folder.

// in the terminal

An engagement runs entirely inside your AI client

No dashboards to open. No PDFs to email. Just a tool call and a folder of artefacts — in whichever client you already use.

co-founder-engine · your AI client
user › run_career_dna("Alex Chen, senior PM at fintech, 8 years")
  run_id: career-dna-alex-chen-2026-04
  Phase 0 of 4 · intake & profile triage
user › get_run_status("career-dna-alex-chen-2026-04")
  Phase 3 of 4 · strategic moves · ETA 4 min
  // 14 minutes later
  Engagement complete · 4 / 4
  → career-dna-snapshot.md  archetype + 3 strategic moves
  → skills-matrix.md        Growth-Share map of skills
  → 90-day-plan.md          Monday-morning actions
  Open report? [yes] / preview / skip
// example output

This is what an engagement delivers

Excerpted from a real Corporate Strategy engagement on a public mid-cap SaaS company. Every number sourced.

co_founder_engine · corporate strategy · 47 min runtime
A data quality
market_segments 7 mapped
Complete
competitors_benchmarked 14 / segment
Strong
hypotheses_tested 10 / 10
All resolved
strategies_generated 82 (12 /seg)
Above bar
data_points_verified 93% (SEC, 10-K)
Excellent
data_points_estimated 6%
Flagged
growth_share_position Star → Cash Cow
Resolved
final-report.md · co_founder_engine output
Verdict
Recommended bet     Vertical SaaS   →  3yr ARR: $48M (P50)
Defend             Mid-market core  →  2 incumbents closing fast
Divest             SMB self-serve  →  CAC payback >36mo
Reasoning trace
H-D1:      “Vertical demand exists” — CONFIRMED
H-A2:      “Distribution moat is durable” — CONFIRMED
H-S1:      “SMB unit economics work” — REJECTED
10/10 hypotheses resolved
93% data verified at source
SEC EDGAR · investor decks · benchmarks
// the difference

Generic AI vs a real engagement

Same prompt: "Should we expand to enterprise?" One answer is opinion. The other is a defensible decision.

− generic AI one-shot answer
## Should we expand to enterprise? Enterprise can offer higher contract values and lower churn. Consider building a sales team and longer implementation cycles. Many SaaS companies have succeeded here. Hope this helps!
opinion No data. No competitor analysis. No hypotheses. Generic platitude — would fail any board review.
+ co-founder engine final-report.md · verdict section
## Verdict: Conditional Go Enterprise expansion: GO — if 2 conditions met by Q3. - H-D1: TAM = $4.1B (Gartner '25) ✅ - H-A2: Channel partner ready ⚠️ - H-S1: CAC payback <18mo ❌ Defer if H-S1 fails by 2026-Q3.
defensible 10 hypotheses tested. Source-cited TAM. Conditional verdict you can present to your board.
// what you actually get

Markdown artefacts, not a deck

Every engagement produces a folder of source-cited, board-ready artefacts. No PDFs to email, no slides to update.

01 / final-report.md

Pyramid Principle verdict

Conclusion first. Three supporting arguments. Source-cited data points. The artefact your board reviews.

Verdict: Conditional Go · H-D1 ✅
H-A2 ⚠️ · H-S1 ❌ · defer if fails Q3
02 / growth-share-matrix.md

Portfolio analysis

Every business unit (or career skill) plotted on the matrix. Stars / Cash Cows / Question Marks / Dogs. Capital allocation logic.

Star: Vertical SaaS · defend & scale
Dog: SMB self-serve · divest by Q3
03 / gtm-playbook.md

Operational GTM plan

ICP, decision-making unit, channel architecture, retention mechanics, 90-day pipeline model. Monday-morning execution.

ICP: Series B vertical SaaS · ARR $5-30M
Channels: outbound + integrator partners
04 / validation-report.md

Source verification

Every number tagged ✅ VERIFIED / ⚠️ ESTIMATED / ❌ NOT FOUND. Data quality score (A / B / C / F). No hallucinations.

Data quality: A (93% verified)
Sources: SEC 10-K, Gartner '25, S&P
// shareable output

Every engagement → public Notion link

Markdown is great for the AI client. A board doesn't open markdown. Each bundle finishes by exporting the full artefact set into a fresh Notion page and returns the public URL alongside the run.

Public URL ready before the model finishes summarising
Engagement page with every artefact as a sub-page
Soft-fail — works without Notion too, you just don't get the URL
No vendor lock-in — raw markdown stays in the run folder
# run_strategy_audit("Acme Inc") response
{
  "run_id": "audit-acme-20260511",
  "status": "completed",
  "eta_minutes": 120,
  "notion_url": "https://www.notion.so/
    Acme-Strategy-Audit-11-05-2026",
  "artefacts": [
    "final-report.md",
    "growth-share-matrix.md",
    "validation-report.md", // 22 more
  ]
}
// engagement types

Eight engagements — from 15-min diagnostic to 4-hour GTM

Pick a fixed-price bundle for the full pipeline, or call any of the 138 atomic tools directly. Free preview — tier gating relaxed until billing launches.

Free15 min

Career DNA

Free 15-min calling-intelligence brief. Try the service before going deeper.

Calling brief + archetype
Skills snapshot
Three highest-EV moves
Notion-shareable URL
Free45 min

Personal Strategy

Operator OS profile + Growth-Share skill map + 3-year roadmap.

Operator OS (P0–P5)
Skills Growth-Share map
8–12 strategic moves
3-year roadmap + 90-day plan
$4960 min

Personal Strategy Deep

Operator OS + outreach pack + salary-negotiation playbook.

Big Five + EQ overlay
Cold-outreach kit
Salary negotiation script
Daily Loop seed plan
Paid2 h

Strategy Audit

Full MBB-grade engagement. 27 agents across 4 phases, fact-checked.

7–15 segments mapped
10 hypotheses tested
Growth-Share matrix + GTM
10–15 strategies / segment
$1994 h

Go-to-Market

Strategy Audit + ICP, DMU, channels, pipeline model, marketing assets.

ICP + decision-making unit
Channel architecture
Google Ads + organic plan
Pipeline model + retention
Paid40 min

Due Diligence

Adversarial investment validation. Red team included by default.

Investment thesis test
Market reality check
Risk matrix (15+ risks)
Go / Conditional / Pass verdict
Subscription~10 min/day

Daily Loop

Today's outreach actions + crafted messages. Real targets, not templates.

3–5 actions for today
Personalised outreach drafts
Pipeline tracking
Weekly review
138 toolsà la carte

Atomic toolkit

Call any single tool directly — map markets, fetch SEC filings, generate Google Ads, score the SRI. list_tools() for the catalogue.

S.L.A.T.E. catalogue
22 categories
Pay-per-call (planned)
Mix & match into custom flows
// the catalogue

152 MCP endpoints in your AI

8 fixed-price bundles for the full pipeline. 138 atomic tools when you want surgical control. 6 utility endpoints for run management. Every endpoint a single agent task — declarative, tested, source-cited.

8 bundles· 138 atomic tools· 6 utility· = 152 endpoints
S
Stakes
What winning means
10 tools
compute_self_realization_index
produce_executive_summary
build_scorecard
L
Landscape
Map reality
35 tools
map_market_segments
fetch_sec_filings
analyze_linkedin_profile
A
Angle
Unfair advantage
27 tools
run_advantage_lens
map_skills_growth_share
profile_operator_drive
T
Trade-offs
Focus by elimination
36 tools
test_dd_hypotheses
build_growth_share_matrix
red_team_thesis
E
Execution
What happens next
28 tools
generate_google_ads
build_corporate_gtm
draft_daily_outreach
5-lens MBB methodology — baked into every engagement Built-in
# Each engagement runs through 5 sequential lenses
- lens_1: Description # what is true today
- lens_2: Advantage # why we can win
- lens_3: Future # 4 forecast scenarios
- lens_4: Options # 10-15 distinct strategies
- lens_5: Selection # capital allocation verdict
# Pyramid Principle output. Every number sourced.
// install — pick your client

Connect in 60 seconds

Co-Founder: Strategy Layer Engine is a remote MCP server. Bring whichever AI client you already use — Claude, ChatGPT, Cursor, Codex, Gemini Code Assist, Cline, Windsurf, anything that speaks MCP. OAuth on first connect, no API keys to manage. Below: the exact steps for each client.

Add Co-Founder Engine to Claude.ai Pro / Max / Team / Enterprise · name + URL prefilled, OAuth on connect
Add to Claude
Claude.ai opens the modal with name and URL prefilled. Confirm and connect — OAuth runs in the same browser. Same flow for Claude Desktop, identical settings screen. // or step-by-step
1
Open Claude Desktop / Claude.ai → Settings → ConnectorsAvailable on Pro, Max, Team, Enterprise plans
2
Click “Add custom connector” → paste the URL
https://mcp.xataco.com/mcp
3
OAuth login opens in browserSign in or register, return to Claude. Tools appear in the side panel.
4
Try it
run a career dna for Alex Chen, senior PM, 8 years
Open ChatGPT connectors Plus / Pro / Team / Enterprise · enable Developer Mode first
Open ChatGPT settings
ChatGPT does not yet support a true one-click MCP install — you'll land on the Connectors panel and paste the URL there. Three clicks total. // step-by-step
1
Settings → Connectors → Advanced → turn on Developer ModeRequired to add custom (non-marketplace) MCP servers
2
Connectors → Create → paste the URL
https://mcp.xataco.com/mcp
3
Authenticate via OAuthBrowser opens, sign in or register, return to ChatGPT. Tools become available across chats.
4
Try it
Use co-founder-engine to run a career DNA for me — senior PM, fintech, 8 years
For Codex CLI / OpenAI Responses API, use the Manual / Other tab with a bearer token.
Single CLI command Claude Code · OAuth on first tool call
// step-by-step
1
Add the server via CLI
claude mcp add co-founder-engine --transport http https://mcp.xataco.com/mcp
2
AuthenticateFirst tool call triggers OAuth in your browser. Token stored in ~/.claude/auth/.
3
Try a slash command
/career-dna John Smith, senior product manager, 8 years experience
Status check anytime: /list_runs or /get_run_status <run_id>
Install in Cursor with one click cursor:// deep-link · OAuth on first call
Add to Cursor
Cursor will open a confirmation dialog. Approve, then OAuth runs in your browser on first tool call. // or manual config
1
Cursor Settings → MCP → Add new MCP serverOr edit ~/.cursor/mcp.json directly
2
Paste this config 
{
  "mcpServers": {
    "co-founder-engine": {
      "url": "https://mcp.xataco.com/mcp",
      "transport": "streamable-http"
    }
  }
}
3
Restart Cursor → OAuth opens on first callTools become available in agent mode and chat.
Install in VS Code with one click vscode:mcp/install · Stable + Insiders
VS Code Insiders
VS Code 1.86+ with GitHub Copilot. Click the button — VS Code opens with a confirm-install prompt. // or manual
1
VS Code → Command Palette → “MCP: Add Server”Requires GitHub Copilot extension with MCP support enabled
2
Choose “HTTP”, paste URL
https://mcp.xataco.com/mcp
3
Approve OAuth flowTools appear in the Copilot Chat tool picker. Use @mcp:co-founder-engine to scope a chat.
1
Cline panel → MCP Servers → Marketplace tab
2
Search “Co-Founder Engine” or add manually 
{
  "mcpServers": {
    "co-founder-engine": {
      "url": "https://mcp.xataco.com/mcp",
      "transport": "streamable-http",
      "alwaysAllow": []
    }
  }
}
3
Save → OAuth on first invocation
1
Windsurf Settings → Cascade → MCP Servers
2
Add server config 
{
  "mcpServers": {
    "co-founder-engine": {
      "serverUrl": "https://mcp.xataco.com/mcp"
    }
  }
}
3
Refresh → OAuth opens in default browser
A
Option A — OAuth (recommended for human users) 
{
  "mcpServers": {
    "co-founder-engine": {
      "url": "https://mcp.xataco.com/mcp",
      "transport": "streamable-http"
    }
  }
}
First request triggers browser OAuth. Token cached locally by your client.
B
Option B — Bearer token (headless / CI / agent-to-agent) 
{
  "mcpServers": {
    "co-founder-engine": {
      "url": "https://mcp.xataco.com/mcp",
      "transport": "streamable-http",
      "headers": {
        "Authorization": "Bearer sk_live_..."
      }
    }
  }
}
Generate keys in your account dashboard. Use Bearer mode for headless automation; OAuth for desktop clients.
C
Verify the connection
curl https://mcp.xataco.com/health
Returns {"status":"healthy","version":"1.0.0"} when the server is up.
D
Available endpoints — 152 in total 8 bundle engagements + 138 atomic tools (S.L.A.T.E. catalogue) + 6 utility endpoints. Bundles run multi-agent pipelines; atomic tools call a single agent for a focused task. Bundles
run_career_dna(name, bio, years_experience) run_personal_strategy(name, bio, years_experience) run_personal_strategy_deep(name, bio, years_experience) run_strategy_audit(company, focus?) run_go_to_market(company, focus?) run_due_diligence(target, thesis?) run_daily_loop(...)
Utility
list_tools() get_run_status(run_id) list_runs(limit?) read_artifact(run_id, filename) enrich_run(run_id, ...) track_action(...)
Atomic tools (sample of 138)
compute_self_realization_index(name) analyze_linkedin_profile(url) fetch_sec_filings(ticker) generate_google_ads(...) map_market_segments(company) ... and 133 more — call list_tools() for the full catalogue
E
Quick notes for other clients Codex CLI / OpenAI Responses API: register the server as a remote MCP tool with the URL above and a bearer token. The tools are auto-discovered through the MCP protocol. Gemini Code Assist: add to ~/.gemini/settings.json under mcpServers using the same JSON shape as Cursor. Goose / Continue / Cody / any other MCP client: the bearer-token JSON above works wherever the client supports HTTP MCP servers. If your client only supports stdio, run a thin local proxy — see docs.
Troubleshooting: OAuth fails → check that your client has popup blockers off and supports OAuth 2.1 + PKCE Connection times out → first cold-start can take ~5s; retry the tool call Tool not visible → restart your client after adding the server config

Every strategic decision without rigor

is a bet you can't justify.

Early access — 43 of 50 spots remaining

Run your first engagement today

Connect your AI client, run a Career DNA, get a Pyramid-Principle report and a 90-day plan. Free. No payment. No lock-in.

Install & run my first engagement →

// MCP connector · OAuth login · no credit card