upseed machine · api / mcp
GTM market intelligence for agents and workflows
Give Codex, Claude, Cursor, and internal agents access to Upseed Machine — company intelligence, signals, fit scores, and evidence from one endpoint. Intelligence only by default.
Returns intelligence only. Does not send emails or make calls on your behalf.
for agents
Built for the tools your team already uses
One MCP server and HTTP endpoint. Drop Machine into any agent or workflow that can call a tool.
Codex
OpenAI agent toolkit
Claude
Claude Desktop + MCP
Cursor
In-editor agents
Internal workflows
Your own pipelines
MCP clients
Any MCP-compatible host
What you get back
Example: a GTM team selling into grocery logistics maps the market, scans target accounts for buying signals, finds entry points, and creates a weekly trigger watch.
Companies
ICP-matched accounts
Signals
Hiring, funding, expansion
Fit scores
0-100 ICP match
Evidence
Why they scored high
Account context
Org, location, structure
Credits used
Transparent metering
Start with a recipe
Most people don't know what to ask an MCP server. Start from a goal — each recipe maps to a chain of tools you can call directly or hand to an agent.
Or call the raw tools
Every capability is a callable tool over HTTP and MCP. Requires a Machine API key issued from Upseed.
Market building
Account intelligence
Always-on monitoring
PARAMETERS
EXAMPLE
curl -X POST https://machine.upseed.ai/api/mcp/execute \
-H "Authorization: Bearer $MACHINE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"tool": "machine_discover",
"args": {
"segment_hint": "precision manufacturing",
"location": "Texas",
"limit": 10
}
}'Chain tools into a workflow
Compose tools into multi-step research your agents can run end to end.
Example workflow
Map grocery distribution in Texas, scan Kroger/H-E-B for middle-mile signals, find entry points, then create a weekly trigger watch.
Response shape
All endpoints return a consistent response structure.
{
"success": true,
"credits_used": 10,
"data": {
"companies": [
{
"domain": "meridianmetal.com",
"name": "Meridian Metal Works",
"score": 87,
"signals": ["hiring_ops", "expansion"],
"employee_count": 180,
"industry": "Precision Manufacturing",
"location": "Houston, TX",
"evidence": [
{
"type": "hiring",
"summary": "Posted 4 ops roles in 30 days",
"confidence": 0.74,
"source": "public web"
}
]
}
],
"total_found": 47,
"returned": 10
}
}Pricing
Private beta pricing is configured by workspace size and usage. Credit tiers below are typical starting points — request access for current ranges.
Starter
500 credits/mo
Core intelligence tools
HTTP + MCP access
Standard rate limits
Pro
2,500 credits/mo
Everything in Starter
+ Market Memory
+ Priority rate limits
+ Trigger Watch monitoring
Enterprise
Custom
Everything in Pro
+ Dedicated infrastructure
+ Custom integrations
+ SLA
Credits roll over for 90 days. Volume discounts available.
Market Memory
Private, persistent learning from your data. Available on Pro and Enterprise plans.
LEARNS FROM
Your customers
Improves lookalike matching over time
Bad fits
Companies you mark as not-a-fit are deprioritized
Saved workflows
Your tool chains become callable units
Run history
Optimizes based on what works for you
ENABLES
Smarter scoring
Scores calibrate to your close patterns
Custom signals
Define what matters for your market
Workflow replay
Re-run saved sequences via API or chat
Persistent context
Agents remember your preferences
Market Memory is private to your account. Your data is never used to train shared models or improve results for other customers.
Request Machine API access
Private beta. We review every request and configure access to your market.
Request API accessReturns intelligence only. Does not send emails or make calls on your behalf.