Onboard a new data source, step 1: create a data spec and get upload URL(s)

First step of setting up a new data integration: creates a data spec. By default (sourceType "file") this returns presigned upload URL(s) for the sample file (a

liveunknown~5s typical

mcpmcp-server

$0.0020

per call

plus $0.001 platform fee per run

Open MCP server; Cracked bills routing only

Provider Upstream docs

$Set up https://cracked-ai-pearl.vercel.app/SKILL.md, then use Cracked to run mcp-com-dpf-it-mcp-server /onboard_data_source for me.

Input

Field	Type	Description
workspaceId	string	Workspace to act on. Defaults to your only workspace if you have exactly one.
specName*	string	Name for the new data spec.
description	string	Optional description of the data spec.
sourceType	string (file \| tables)	Defaults to "file" (upload a sample file). Use "tables" to query existing workspace table(
sourceTables	array	sourceType "tables" only: names of existing workspace tables the generated query reads fro
targetOption	string (auto-infer \| existing-tables \| target-schema-file)	Where transformed data should land — works the same for both sourceType values: "auto-infe
targetTables	array	Names of existing workspace tables to target — exactly one entry for sourceType "tables" (
targetSchemaFileName	string	File name of a target schema file. Required when targetOption is "target-schema-file", for
sampleFileName	string	sourceType "file" only (and required for it). File name of the sample data file (e.g. "cus
formatFileName	string	sourceType "file" only. File name of an optional format spec file.
merge	boolean	Upsert instead of plain append when true (default false). For sourceType "tables": generat
additionalPrompt	string	Instructions for the AI. For sourceType "tables", describe what the query should compute f
autoRefresh	string (spec_success \| schedule \| none)	sourceType "tables" only. Required for it — ask the user rather than assuming, and do not
autoRefreshUpstreamSpecName	string	Required when autoRefresh is "spec_success". The spec whose successful job completion shou
autoRefreshFrequency	object	Required when autoRefresh is "schedule".

Call it

curl

curl https://cracked-ai-pearl.vercel.app/v1/run \
  -H "Authorization: Bearer ck_live_..." -H "content-type: application/json" \
  -d '{"provider":"mcp-com-dpf-it-mcp-server","endpoint":"/onboard_data_source","input":{"specName":"..."}}'

cli

npx cracked-ai run -p mcp-com-dpf-it-mcp-server -e /onboard_data_source -i '{"specName":"..."}'

mcp

run_tool({ provider: "mcp-com-dpf-it-mcp-server", endpoint: "/onboard_data_source", input: {"specName":"..."} })

Try it

Runs against your signed-in workspace balance. Sign in if you have not.