Key facts
| Vision input | Image understanding runs on the OpenAI-compatible chat completions endpoint with supported multimodal models |
| Example model | plugsky-gemma-4 handles text plus image inputs such as screenshots, forms and product photos |
| Model access | 30+ models from efficient to frontier tiers behind one key |
| Dedicated media endpoints | Image generation and editing endpoints are labelled coming soon |
| Deployment | Hosted, VPC, on-prem and air-gapped options on Enterprise |
| Data handling | Data encrypted at rest and in transit; API data is not used to train models |
| Audit | Per-request logs with model, tokens, latency and user; SIEM export; retention up to 7 years |
| Free plan | plugsky-micro and plugsky-lite, no card; 14-day full-access trial |
TL;DR
- Extract structured fields from forms and permits; keep officers on exceptions.
- Process citizens’ images in a region-locked, on-prem or air-gapped deployment.
- BYOK keeps encryption keys with the institution.
- Audit export with retention up to 7 years supports records and reviews.
- Automated extraction is evidence — decisions keep human review.
How it works, step by step
- Choose a non-sensitive, high-volume form as the first workload.
- Define the extracted fields and the exact-match accuracy target.
- Run a labelled sample through the OpenAI-compatible vision call with a JSON schema.
- Keep officers reviewing exceptions and low-confidence fields.
- Select the deployment plane and key custody that match the data classification.
- Export audit events to the departmental SIEM with policy-matched retention.
- Extend to higher-sensitivity documents only after the pilot is signed off.
Original data
Try it yourself
Open the model availability tracker →
Why government need image understanding
Government document processing is high-volume, rule-bound and auditable: licence applications, permit photos, identity documents, inspection reports. Staff verify fields one by one, and backlogs grow. Vision models can extract those fields into a schema, leaving officers to handle judgment calls and exceptions rather than transcription.
Two constraints shape the design. Citizen images are personal data, so processing location and retention must be defensible. And decisions with legal effect need human review and a record — an automated extraction is evidence, not an adjudication.
How vision works on the Plugsky API
Image understanding does not need a new platform. Add the image to the message content of a standard chat completion, and the response arrives exactly like a text completion: streamable, validatable with JSON mode, and usable with function calling. plugsky-gemma-4 covers high-volume triage such as screenshots, forms and document images, while stronger models handle the hard cases. Because 30+ models sit behind one key, changing tier is a model-name change. For government, the same call shape works across region-locked, on-prem and air-gapped planes, so classification decides placement, not code.
The controls government should configure first
Sovereign delivery is deployment-based. Region-locked planes keep processing in a chosen geography; VPC, on-prem and air-gapped installations cover programmes that cannot use a shared environment. Air-gapped deployments run with no internet egress, a local model registry and offline update channels, so images never leave the perimeter.
Keys stay with the institution through BYOK with AWS KMS, Azure Key Vault, HashiCorp Vault or an on-prem HSM. Per-request audit logs record model, tokens, latency, user and region and export to your SIEM with retention up to 7 years, supporting freedom-of-information, records and review processes. Note that image generation endpoints are labelled coming soon; this architecture is about understanding documents, not producing them.
A practical pilot path
Pilot on a non-sensitive, high-volume form with clear fields and an existing manual baseline. Measure field accuracy and exception rates, keep officers in the loop, and only then move to documents with higher sensitivity. Prototype on the free plan, then deploy in the plane the programme requires. See the live pricing page for current plans.
Honest comparison
| Capability | Plugsky | Vision-only API vendor | Building in-house |
|---|---|---|---|
| Image understanding | Live on supported multimodal chat models | Usually the core product | You host and serve vision models |
| Image generation | Labelled coming soon | Often available today | Custom deployment work |
| API shape | OpenAI-compatible chat completions | Vendor-specific SDK | You standardise it yourself |
| Model choice | 30+ models behind one key | Narrower catalogue | One model per integration |
| Deployment | Hosted, VPC, on-prem and air-gapped | Usually hosted only | Your infrastructure only |
| Controls | Region pinning, BYOK, per-request audit logs, SIEM export | Varies by vendor | You build the pipeline |
Frequently asked questions
Can vision run in an air-gapped environment?
Yes, on supported open-weight multimodal models in an air-gapped deployment with no internet egress, a local model registry and offline update channels.
Do images leave our network?
In hosted or region-locked deployments they go to the selected data plane only; in air-gapped deployments nothing leaves the perimeter. Deployment choice is the control.
Do we have to change our application code?
No. Image understanding uses the same OpenAI-compatible chat completions endpoint, so your SDK, auth and error handling carry over; the message content simply includes an image.
Is there a free plan?
Yes — plugsky-micro and plugsky-lite with 2 API keys and no credit card. A 14-day full-access trial is also available.
How is pricing structured?
Self-serve plans are flat monthly with unlimited fair-use usage and no per-token charges or overage fees. See the live pricing page for current plans.
Is image generation available?
Not yet — image generation and editing endpoints are labelled coming soon. Image understanding is live today on supported multimodal chat models.