Treasure Data

REST API at api.treasuredata.com. Resources: databases, tables, jobs, bulk_imports, segment_configs. GET /v3/database/list lists databases. POST /v3/table/import/{db}/{table}/json imports JSON records (real-time ingestion). POST /v3/job/issue/hive/{db} runs a Hive query; POST /v3/job/issue/presto/{db} runs a Presto (Trino) query. GET /v3/job/status/{job_id} polls async query status. GET /v3/job/result/{job_id} retrieves query results (msgpack or JSON). Real-time event ingestion at api.in.treasuredata.com/js/v3/event/css/{db}/{table} (HTTP event collector). Bulk import: td-agent (Fluentd plugin) or td CLI.

Standard HTTP status codes with JSON or msgpack error bodies. 401 for invalid API key, 409 for duplicate import job, 500 for server error. Async query jobs: poll /v3/job/status until status is 'success' or 'error'. Query result download via /v3/job/result as msgpack (efficient) or JSON. No client-facing rate limits documented (bulk ingestion handles high volumes). Treasure Data platform uptime documented at status.treasuredata.com. Data pipeline SLA for enterprise accounts.

Treasure Data is an enterprise CDP — less common in startup stacks, but deeply embedded in retail, automotive, and telco enterprises (Toyota, SoftBank, major retailers). For agents operating in those environments, it may be the authoritative customer data spine. Capabilities: real-time event ingestion, SQL-queryable customer profiles (Presto/Hive), audience segmentation, and activation to downstream channels. REST API for data management; bulk import via td-agent (Fluentd-based) or direct import API. Self-serve setup is limited — enterprise account required. Confidence is docs-derived; lower confidence due to enterprise access barriers.

API key auth: TD1 {api_key} header on all REST API requests. API keys from Treasure Data console → Account Settings. Single API key per account (no per-project scoping at base tier). HTTPS enforced. IP allowlisting available for enterprise accounts. No OAuth2. Write-only API keys available for client-side event collection. Enterprise RBAC for multi-user account management.

docs.treasuredata.com provides REST API reference, data ingestion guide (event collection API, bulk import, td-agent setup), SQL query engine documentation (Presto/Hive), audience segmentation guide, and activation connectors. Documentation is comprehensive but assumes enterprise context. Getting started requires a Treasure Data account (enterprise trial or production account). td CLI (Ruby gem) for command-line management. Python SDK available (pytd). Community support via Treasure Data support portal (enterprise SLA).