Recommend logo
Book demo
Agent API

Connect your agents to live commercial context

One call over MCP or HTTP returns your processed data and executes the action.

Book demo
Connects to your stack
ShopifySalesforceGoogle TrendsMetaTikTokGA4AmazonStripePricing / ERPCRMEditorial feeds
How to use agents

One call. Full context. Executed action.

01 · Key

Get a key

Create a key in the dashboard and export it. That’s the only setup there is.

shell
# create a key in the dashboard, then:
export RECOMMEND_API_KEY=rcmd_live_...
02 · Connect

Register the MCP server

One config block and your agent calls live context natively — no glue code, no SDK to learn.

mcp.json
{
  "mcpServers": {
    "recommend": {
      "url": "https://mcp.recommend.co",
      "apiKey": "rcmd_..."
    }
  }
}
03 · Execute

Execute the move

Your agent reads the live context and acts on it — every action logged, every outcome feeding back into the next run.

your agent · mcp
get_context({ object: "sku:4812", market: "DE" })
→ { "demand": 78, "move": "push_in_DE" }

act({ move: "push_in_DE" })
→ brief drafted · budget +15% · meta live
Tools

What your agents can call

get_context()mcp · rest

The live state of any object — demand, stock, price position, competitors, and the prepared move.

act()mcp · rest

Execute the prepared move: draft the brief, shift the budget, launch the campaign.

query()mcp · rest

Ask in natural language and get grounded, current numbers back.

watchlist()mcp · rest

Every object your team follows — the same objects, the same states.

outcomes()mcp · rest

Settled results, and what each one recalibrated.

calibration()mcp · rest

The track record behind every score.

From context to execution.

Book a demo and watch your data go from context to executed action.

Book demo