Introduction

Support DQE
Support DQE
  • Updated
Introduction - DQE One Standalone API

The DQE One Standalone API lets you control your self-hosted DQE One Standalone server remotely: upload files, connect data sources, create and run data-quality processes, follow their runs and manage RuleSets — everything you can do in the web interface to automate batch processing.

{DOMAIN} is the host name of your own DQE One Standalone server — it is self-hosted, so this value depends on your deployment. {EMAIL} is your DQE One login and {API_TOKEN} the token generated in My Account > API Token.
Not covered here: the real-time duplicate lookup API (Use the Lookup API in DQE One Standalone) and the MCP server for AI assistants (DQE One Standalone — MCP server).
Jump toAuthenticationConventionsAPI mapEnd-to-end exampleLimitations

Authentication

Every endpoint uses HTTP Basic authentication: your DQE One email as the username and your API token as the password (Authorization: Basic base64(email:api_token); with curl, -u "email:token"). No session, Bearer token or OAuth is accepted on these routes.

  1. Sign in to your DQE One Standalone web interface.
  2. Open My Account and the API Token section, then generate the token.
  3. Use it as the Basic password. Regenerating it invalidates the previous one.
Treat the token like a password. It never expires, has no scopes and gives full access to everything your account owns (files, credentials metadata, processes, RuleSets). Keep it out of source control and logs. See Manage your profile in DQE One Standalone for how to renew it.

Conventions

Topic Behaviour
Base URL https://{DOMAIN}/api/... — DQE One Standalone is self-hosted, so {DOMAIN} is your own server.
Ownership Everything is scoped to the authenticated user. Another user's id answers 404, never 403.
Success {"status":"success", ...}. Lists are bare JSON arrays. Creation endpoints answer 201, including updateProcess.
Errors {"status":"error","error":"<message>"} in English. Exceptions are flagged on each endpoint (for example runProcess uses message, and RuleSet import returns error as a list).
401 {"status":"error","error":"Unauthorized"} for a missing, wrong or unknown credential. A malformed header (not base64, or without :) can produce an unformatted 500 instead.
Methods A wrong method answers 405. runProcess and deleteProcess are GET requests that change data.
Asynchronous work runProcess only queues the run. Poll getRuns. Jobs run one at a time per user.
Unformatted 500 A few inputs (a non-numeric id, a process that never ran) raise a plain server error instead of JSON. Validate ids on your side.
Limits No rate limiting, no pagination, no explicit upload size limit, tokens never expire. CORS is open to any origin (authority comes from the Authorization header).
Interactive docs Your server also serves ReDoc at /doc/ and Swagger UI at /doc/swagger/. They are not fully accurate (key names, status codes, array vs object responses): prefer this documentation and the OpenAPI file above.

API map

Article Endpoints
Files addFile, listFiles, getProcessResult, deleteFile
Credentials createCredential, listCredentials, getCredential, deleteCredential
Processes createProcess, updateProcess, listProcesses, getProcess, runProcess, deleteProcess
Runs & results getRuns
RuleSets listRulesets, getRuleset, exportRulesets, importRuleset
MCP server The same operations for AI assistants (Claude and other MCP clients): JSON-RPC at /mcp.

Looking for real-time duplicate lookup? That is a separate API: Use the Lookup API in DQE One Standalone.

End-to-end example

This complete flow needs nothing but a CSV file. For a database or CRM source, create a credential first (Credentials) and use "source":"postgres", "object":"<table>", "credential":<id> instead of file.

export DOMAIN=standalone.example.com EMAIL=you@example.com TOKEN=your_api_token # 1. Upload a file (headers: unique, no spaces) curl -u "$EMAIL:$TOKEN" -F "file=@contacts.csv" https://$DOMAIN/api/File/add # - 201 {"status":"success","id":12} # 2. Create a Check Email process on it curl -u "$EMAIL:$TOKEN" -H "Content-Type: application/json" \ -d '{"name":"Email check","kind":"check_email","source":"csv","file":12,"primaryKey":"Id","fields":{"FieldLabel":"Email"}}' \ https://$DOMAIN/api/Process/create # - 201 {"status":"success","id":42} # 3. Run it (asynchronous) curl -u "$EMAIL:$TOKEN" "https://$DOMAIN/api/Process/run?id=42" # - 200 {"status":"success","message":"Process Enqueue","run_id":"101"} # 4. Poll until status is Finished or Failed curl -u "$EMAIL:$TOKEN" "https://$DOMAIN/api/Run/get?id=101" # 5. Download the result (primary key + synchronised columns) curl -u "$EMAIL:$TOKEN" -OJ "https://$DOMAIN/api/File/get?processId=42"

Clean-up order matters: delete the process first (once its run is Finished or Failed), then the credential, then the file; each is refused with 423 while something still uses it.

Known limitations

  • find_and_replace, format and audit processes cannot be created reliably through the API, and Snowflake processes always fail validation: use the web interface for these.
  • updateProcess cannot set a boolean back to false and resets the fields you do not send.
  • There is no endpoint to cancel a run, and getProcessResult returns only the key and synchronised columns.
  • Dynamics and Snowflake credentials need a browser login to complete.

These describe the current behaviour of the product, not planned changes.

See also

Related to

Was this article helpful?

0 out of 0 found this helpful