OpenAI Providers Update - July 2026
We've released an update to the StackQL providers for the OpenAI platform:
openai- the platform API surface available to standard API keys: models, files, fine-tuning, batches, vector stores, assistants, evals, conversations, uploads, containers, and skills (11 services, 26 resources, 97 operations)openai_admin[new] - the organization and administration API surface: usage and cost reporting, projects, organization users and invites, groups and roles, admin API keys, audit logs, and certificates (10 services, 29 resources, 81 operations)
Both providers expose a SQL-first surface: authentication is handled automatically, push down support using the LIMIT clause and built in pagination handling.
The openai provider is a ground-up rebuild of the previous provider, generated from the vendor's published OpenAPI specification. Some resources have been renamed and the organization/admin surface has moved to openai_admin - the previous provider version remains available in the registry for pinning, and the full disposition is documented at openai-provider.stackql.io.
Async jobs as SQL
Fine-tuning jobs, batches, vector store file batches and uploads follow the same pattern: INSERT creates the job, SELECT polls it, EXEC cancels it. For example:
SELECT id, status, model, fine_tuned_model, trained_tokens
FROM openai.fine_tuning.jobs;
SELECT id, status, endpoint, request_counts
FROM openai.batches.batches
LIMIT 10;
Vector stores are a full CRUD surface with file membership:
SELECT id, name, status, usage_bytes, file_counts
FROM openai.vector_stores.vector_stores;
Inference endpoints (chat/completions, responses, embeddings, images, audio) are deliberately out of scope - the provider covers the control plane; use the vendor SDKs for invocation.
The admin provider - your OpenAI org as data
The openai_admin provider presents the organization management surface as SQL.
Usage and cost are bucketed time series: one row per time bucket, with the per-group breakdown in the results JSON column, fanned out with JSON_EACH. Token usage by project and model over a 30-day window:
SELECT
json_extract(r.value, '$.project_id') AS project_id,
json_extract(r.value, '$.model') AS model,
strftime('%Y-%m-%d', u.start_time, 'unixepoch') AS usage_date,
json_extract(r.value, '$.input_tokens') AS input_tokens,
json_extract(r.value, '$.output_tokens') AS output_tokens
FROM openai_admin.usage.completions u, json_each(u.results) r
WHERE u.start_time = 1781481600
AND u.bucket_width = '1d'
AND u.limit = 31
AND u.group_by = 'project_id'
ORDER BY usage_date, project_id;
Daily spend in USD by project:
SELECT
strftime('%Y-%m-%d', c.start_time, 'unixepoch') AS cost_date,
json_extract(r.value, '$.project_id') AS project_id,
json_extract(r.value, '$.amount.value') AS amount_usd
FROM openai_admin.costs.costs c, json_each(c.results) r
WHERE c.start_time = 1781481600
AND c.limit = 180
AND c.group_by = 'project_id'
ORDER BY cost_date, amount_usd DESC;
Usage is broken out per capability - completions, embeddings, moderations, images, audio, vector stores and code interpreter sessions - and can be grouped by project_id, api_key_id or model.
Governance and auditing
Projects and their child resources (users, service accounts, API keys, rate limits) are queryable and writable:
SELECT p.name AS project, p.status, sa.name AS service_account, sa.role
FROM openai_admin.projects.projects p
JOIN openai_admin.projects.service_accounts sa ON sa.project_id = p.id
WHERE p.status = 'active';
Admin key hygiene and audit logs work the same way:
SELECT name, created_at, last_used_at, owner
FROM openai_admin.admin_api_keys.admin_api_keys
ORDER BY created_at;
SELECT id, type, effective_at, actor, project
FROM openai_admin.audit_logs.audit_logs
WHERE "effective_at[gt]" = 1750000000
AND "event_types[]" = 'project.created';
Because it's all SQL, you can join usage to projects, materialize daily cost snapshots into a database, or point a BI tool at StackQL's Postgres wire protocol server and build an org-wide OpenAI spend dashboard without writing a line of integration code.
Authentication
The two providers use different key types, which are disjoint by design:
# openai - standard API key
export OPENAI_API_KEY=sk-...
# openai_admin - org-scoped admin key (created by organization owners)
export OPENAI_ADMIN_KEY=sk-admin-...
Admin keys are available to organization accounts only and can be provisioned by organization owners in the platform console. A standard key cannot call the admin endpoints and vice versa.
Get started
Pull the providers from the public registry:
registry pull openai
registry pull openai_admin
Provider docs are at openai-provider.stackql.io and openai-admin-provider.stackql.io. Let us know what you build. Star us on GitHub.