Image visual question answering

POST /api/image-vqa10 quota units per call · cache hits free

Ask LLaVA a question about an image and receive a natural-language answer. Powered by Cloudflare Workers AI. Useful for content extraction, visual reasoning, and accessibility.

Pose a free-form question about the input image and get a natural-language answer from the LLaVA 1.5 vision-language model. Sibling of `image-caption` — same model, different prompt shape. Useful for: - Extracting a specific value from a photo ("What's the price on this receipt?", "How many people are in the photo?"). - Yes/no triage ("Does this image contain text?"). - Accessibility-side queries that go beyond a generic caption. **Cost: paid Cloudflare Workers AI inference.** Same model as image-caption — charges against your account's Workers AI budget plus 10 wasm-tool quota units per call.

Inputs

Name	Type	Default	Description
image*	file	—	Image to ask about (PNG, JPEG, WebP, AVIF).
question*	string	—	The question to ask. Phrase as a direct question — e.g. "What's the price on this receipt?".
maxTokens	number (16…1024)	256	Maximum tokens in the answer.

Response

Modes: json. Cache: yes (24h TTL).

Every response carries x-cache (HIT / MISS / BYPASS), x-cache-signature (stable across identical inputs), and the rate-limit headers listed below.

Quota & limits

Cost per call	10 units against the 10,000-unit monthly quota. Cache hits are free — only cache misses decrement.
Burst limit	200 requests per rolling minute, shared across all tools.
Max request size	`4 MiB` (base64 inflates file payloads ~33% — the limit applies to the decoded bytes).
Max response size	`16 KiB`
Timeout	60s wall clock.
Rate-limit headers	`X-RateLimit-Limit`, `X-RateLimit-Remaining`, `X-RateLimit-Burst-Limit` on every response; `X-Quota-Warning` once 80% of the monthly quota is spent; `Retry-After` on 429s.

Errors

Errors are JSON with an error string; schema failures add a details object (Zod's flattened field errors).

Status	When	Body
400	Input fails the tool's schema (wrong type, missing required field, malformed base64) or the file bytes can't be parsed as the expected format.	{ "error": "invalid inputs", "details": { "fieldErrors": { … } } }
401	Missing or invalid X-Api-Key / X-Account-Id headers (or no session when called from the browser).	{ "error": "unauthorized" }
404	Unknown tool id, or calling /api/* on the website hostname instead of api.xtract.bot.	{ "error": "not found" }
413	Request body over the tool's max request size, or an image header declaring more megapixels than the tool's pixel budget.	{ "error": "…exceeds maxRequestBytes…" }
415	Content-Type isn't application/json on the json-body transport.	{ "error": "unsupported content-type; expected application/json" }
429	Monthly quota or the 200/min burst limit exhausted. Check Retry-After and the X-RateLimit-* headers.	{ "error": "monthly quota exceeded", "monthRemaining": 0, … }
5xx	Conversion engine failure. Safe to retry; if it persists, the input is hitting a bug — please report it.	{ "error": "…" }

Code samples

Built from the what-is-this example.

# Download or substitute the example input:
#   curl -O https://xtract.bot/examples/image-align/sample.png
IMAGE=$(base64 -w0 < sample.png)

curl -X POST https://api.xtract.bot/api/image-vqa \
  -H "Content-Type: application/json" \
  -H "Accept: application/json" \
  -H "X-Account-Id: $XTRACT_ACCOUNT_ID" \
  -H "X-Api-Key: $XTRACT_API_KEY" \
  -d '{
  "question": "What is shown in this image?",
  "image": "'"$IMAGE"'"
}'