The customer-context layer
for AI agents and humans.

One layer in the middle. Anything on either side.

Any MCP-compatible client or REST caller on top. Practically anything that holds customer data on the bottom. Pathbound is the layer that makes them speak the same language.

Upstream · 01 Anything that speaks MCP or REST

Claude ChatGPT Gemini Grok Mistral Perplexity ⌘ Cursor Claude ChatGPT Gemini Grok Mistral Perplexity ⌘ Cursor

n8n n8n {} LangChain ▮▯ Vercel AI SDK ⚙ Zapier Agents ⌥ Raycast AI </> Custom backend n8n n8n {} LangChain ▮▯ Vercel AI SDK ⚙ Zapier Agents ⌥ Raycast AI </> Custom backend + any MCP client + any MCP client

Pathbound

the customer-context layer

01 Collect 02 Resolve 03 Prepare 04 Protect

Downstream · 02 Connect via OAuth, webhooks, JS tracker

HubSpot Salesforce Pipedrive Zoho CRM Intercom Zendesk Gmail Resend SendGrid Mailchimp HubSpot Salesforce Pipedrive Zoho CRM Intercom Zendesk Gmail Resend SendGrid Mailchimp

Apollo LinkedIn Slack Stripe Chargebee Postgres BigQuery Zapier Webhooks Pathbound Tracker Apollo LinkedIn Slack Stripe Chargebee Postgres BigQuery Zapier Webhooks Pathbound Tracker + 40 more integrations + 40 more integrations

// Switch the agent stack later — or add a new CRM tomorrow — without rebuilding the pipeline. The layer in the middle stays the same.

From silo to LLM, in four stages.

Two surfaces: MCP and REST.

Same data, same identity resolution, same response shapes — pick whichever fits your stack.

{
  "mcpServers": {
    "pathbound": {
      "url": "https://mcp.pathbound.ai",
      "headers": {
        "Authorization": "Bearer sk_live_..."
      }
    }
  }
}

curl "https://api.pathbound.ai/v1/contacts/events/ctc_01H..." \
  -H "Authorization: Bearer sk_live_..."

const r = await fetch(
  `https://api.pathbound.ai/v1/contacts/events/${id}`,
  { headers: { Authorization: `Bearer ${process.env.PATHBOUND_API_KEY}` } }
);
const timeline = await r.json();

import os, requests
r = requests.get(
    f"https://api.pathbound.ai/v1/contacts/events/{id}",
    headers={"Authorization": f"Bearer {os.environ['PATHBOUND_API_KEY']}"},
)
timeline = r.json()

{
  "contact_id": "ctc_01H...",
  "email": "[email protected]",
  "company": { "domain": "acme.com", "industry": "fintech", "employees": 240 },
  "lifecycle_stage": "trial",
  "last_seen_at": "2026-04-29T14:02:11Z",
  "summary": "Senior PM evaluating Acme's billing integration. Replied to onboarding yesterday; visited /pricing 2× today.",
  "events": [
    { "type": "page_view",         "source": "tracker",  "at": "…14:02:08Z" },
    { "type": "form_submission",   "source": "tracker",  "at": "…14:01:42Z" },
    { "type": "email_reply",       "source": "gmail",    "at": "…09:30:00Z" },
    { "type": "deal_stage_change", "source": "hubspot",  "at": "…16:00:00Z" }
  ],
  "_meta": { "tokens": 312, "pii_redacted": false, "freshness_ms": 480 }
}

Every event from your website and app — tied back to a real customer, even the ones from before they signed up.

The pricing-page visits, the half-watched demo, the blog posts that brought them in. All of it stitched onto the contact the moment they sign up — so your agent can answer "why is Jane here, and how warm is she?" with real evidence.

{ "contact_id": "ctc_01H…", "email": "[email protected]", "name": "Jane Reyes", "title": "Senior PM",
  "company": { "domain": "acme.com", "industry": "fintech", "employees": 240 },
  "lifecycle_stage": "trial", "intent_score": 0.86, "first_seen_at": "2026-04-15T…",
  "summary": "Senior PM evaluating billing integration. /pricing 3×, demo 75%, read Atlas case study. High intent, late-funnel.",
  "signals": ["pricing_intent", "compared_alternatives", "technical_evaluator"],
  "events": [
    { "type": "page_view", "path": "/blog/customer-context", "at": "…14d", "resolved_from": "visitor_a91c" },
    { "type": "page_view", "path": "/pricing",                "at": "…6d",  "scroll_depth": 1.0 },
    { "type": "video",     "name": "product_demo",         "at": "…6d",  "progress": 0.75 },
    { "type": "page_view", "path": "/pricing",                "at": "…2d",  "visit_count": 2 },
    { "type": "signup",    "email": "[email protected]",        "at": "…now", "merged_visitors": 1 }
  ],
  "_meta": { "tokens": 312, "pii_redacted": false, "freshness_ms": 480 }
}

Drop a script. Set a cookie. That's it.

No SDK to install, no .identify() to call, no event taxonomy to design. The tracker auto-captures what matters; the cookie is the identity signal.

<script async src="https://tracker.pathbound.ai/tracker.js"></script>

// In your app, after login or signup:
document.cookie =
  "pathbound_external_contact_id=user_4f3a...;"
  + " path=/; max-age=31536000; SameSite=Strict; Secure";

DNT honored · HTTPS-only · no third-party cookies · content-stable fingerprint (not cross-site identifying) · full audit trail of every merge.