Skip to content

Repository files navigation

Vision API — Go client

Official Go client for Vision API — send an image or a PDF, describe the fields you want in plain language, get structured JSON back with a confidence level on every value.

Go Reference Go Report Card license


Install

go get github.com/devrobotlabs/visionapi-go

Go 1.21+. Standard library only — no dependencies to vendor or audit.

Quick start

package main

import (
	"context"
	"fmt"
	"log"

	"github.com/devrobotlabs/visionapi-go"
)

func main() {
	client, err := visionapi.New() // reads $VISION_API_KEY
	if err != nil {
		log.Fatal(err)
	}

	res, err := client.Analyze(context.Background(), visionapi.AnalyzeParams{
		Source: visionapi.FilePath("invoice.pdf"),
		Preset: "invoice",
	})
	if err != nil {
		log.Fatal(err)
	}

	fmt.Println(res.Result.Field("invoice_id").String()) // "A-10422"
	total, ok := res.Result.Field("total").Float()       // ok is false when absent
	fmt.Println(total, ok, res.CreditsUsed, res.CreditsRemaining)
}

Requests are metered in credits, per image and per selected PDF page — see pricing for current rates. Failures cost nothing: the reservation is released in full on any non-2xx, so there is no compensating logic to write.

Server-side only. There is no publishable key and no test mode — an API key is a live spending credential. Never embed one in a client binary you ship.


Reading a result

Two rules explain almost every surprise:

1. Every scalar is wrapped in {value, confidence}, where confidence is Low, Mid or High.

2. A preset response contains every field of that preset — including the ones the document does not carry, which come back with a nil value. A key being present does not mean a value was found.

Result gives you typed accessors that are safe on any response, including for a field that isn't there:

r := res.Result

r.Field("invoice_id").String()          // "" when absent
r.Field("total").Float()                // (float64, bool)
r.Field("is_paid").Bool()               // (bool, bool)
r.Field("invoice_date").Date()          // (time.Time, bool) — the API normalizes to ISO 8601
r.Field("reference_numbers").Strings()  // []string for an array-of-scalars field
r.Field("carrier").IsNull()             // true — present in the preset, absent from the document
r.Field("total").Confidence.AtLeast(visionapi.High)

Line-item arrays are the one shape worth looking at twice. The array itself is not wrapped; each cell inside each row is:

for _, row := range r.Rows("line_item") {
	fmt.Println(
		row.Field("description").String(),
		row.Field("quantity").Int(),
		row.Field("amount").Float(),
	)
	// or row.Plain() for a map[string]any ready for a CSV writer
}

An object field works the same way: r.Block("seller").Field("name").String().


What you can send

Exactly one source per call:

visionapi.FilePath("invoice.pdf")                    // read from disk
visionapi.FileBytes("scan.png", data)                // bytes you already hold
visionapi.FileReader("scan.png", r)                  // any io.Reader (buffered, so retries resend)
visionapi.FileURL("https://example.com/invoice.pdf") // the API fetches it
visionapi.FileBase64(encoded)                        // "data:" prefix optional

JPEG, PNG, WebP, TIFF and PDF, up to 20 MB and 50 pages. The type is detected from magic bytes — the filename is ignored.

AnalyzeParams

Field Default What it does
Preset A catalog name, or "auto" to let the API classify the file first (free).
Schema Custom fields, alone or on top of a preset.
SchemaName A schema saved in your dashboard. Excludes Preset and Schema.
Pages all PDF page selection, e.g. "1-3,7". You pay for selected pages only.
LanguageHint auto ISO 639-1 code, e.g. "es".
Detail "standard" "high" renders pages at higher resolution. Same cost, slower.
Output "json" "text" returns raw OCR text instead of fields.
IncludeRawText false Adds FullText, the whole transcription, alongside Result.
MinConfidence Low Fields below the level come back null, with confidence preserved.

Custom fields

A schema is a flat map: each key is a field name, each value describes what to extract. It is compiled before any credit moves, so a bad schema costs nothing.

res, err := client.Analyze(ctx, visionapi.AnalyzeParams{
	Source: visionapi.FilePath("invoice.pdf"),
	Preset: "invoice",
	Schema: visionapi.Schema{
		// Plain form — the string is the description, type defaults to string.
		"machine_serial": `Serial number of the machine being invoiced, without the "SN:" prefix`,

		// Typed form.
		"total_net": visionapi.FieldSpec{Type: "number", Description: "Total before tax"},
		"signed_on": visionapi.FieldSpec{Type: "date", Description: "Date the contract was signed"},

		// Reserved key: injects fields into every row of the preset's line-item array.
		"line_item": visionapi.Schema{"lot_number": "The lot number printed on the line"},
	},
})

Field names must match ^[a-z][a-z0-9_]{0,63}$. Types are string (default), number, boolean, date, array and object. A custom name that collides with a preset field is a 422 schema_field_conflict — rename it, or use the preset's own field.

Descriptions are the prompt. "The invoice number exactly as printed, without the #" extracts better than "invoice number". Say what to do when the value is missing or ambiguous if it matters.

Reuse a combination by saving it:

client.CreateSchema(ctx, "our-invoices", "invoice", schema)
client.Analyze(ctx, visionapi.AnalyzeParams{Source: src, SchemaName: "our-invoices"})

Picking a preset

28 presets ship with the API. Fetch the catalog rather than hardcoding field names from memory — presets are versioned, and the catalog is the source of truth:

presets, _ := client.Presets(ctx)          // no API key required
invoice, _ := client.Preset(ctx, "invoice")
for _, f := range invoice.Fields {
	fmt.Println(f.Name, f.Type, f.Description)
}

Three ways to choose:

// 1. You know what it is.
client.Analyze(ctx, visionapi.AnalyzeParams{Source: src, Preset: "receipt"})

// 2. You don't, and you want the data anyway. Classification is free.
res, _ := client.Analyze(ctx, visionapi.AnalyzeParams{Source: src, Preset: "auto"})
res.Detection.Preset       // what ran
res.Detection.Fallback     // true = "shape unknown", not a match
res.Detection.Alternatives // the rest of the ranking, best first

// 3. The *type* is the decision — routing a mixed inbox, or refusing to spend
//    on a 40-page PDF until you know what it is. Far cheaper than extracting.
guess, _ := client.Detect(ctx, visionapi.DetectParams{Source: src})
if guess.Recommended == "invoice" && !guess.Fallback {
	client.Analyze(ctx, visionapi.AnalyzeParams{Source: src, Preset: "invoice"})
}

Detect reads page 1 only, so an image and a 300-page PDF cost the same, and it is metered in batches rather than per call: most calls report CreditsUsed: 0 and an occasional one carries the charge. See pricing for the rate.


Questions instead of fields

Up to 5 questions about one file, priced exactly like an extraction. The questions themselves are free.

res, err := client.Ask(ctx, visionapi.AskParams{
	Source:    visionapi.FilePath("photo.jpg"),
	Questions: []string{"Is there a dog in the image?", "How many people are visible?"},
})

for _, a := range res.Answers {
	switch a.Verdict {
	case visionapi.Yes:
		// …
	case visionapi.Uncertain:
		// A yes/no question the image does not settle — a real answer, not a failure.
	case visionapi.NotApplicable:
		// It wasn't a yes/no question; read a.Answer.
	}
}

Long jobs: async and webhooks

Synchronous requests are killed at 60 seconds with a 504 sync_timeout. Anything that might run longer — a long PDF, Detail: "high", a batch — belongs on the queue.

// Submit, then poll. WaitForTask handles the loop and the failure case.
task, err := client.AnalyzeAndWait(ctx,
	visionapi.AnalyzeParams{
		Source: visionapi.FilePath("contract-80-pages.pdf"),
		Preset: "contract",
		Pages:  "1-50",
	},
	visionapi.PollInterval(2*time.Second),
	visionapi.MaxWait(15*time.Minute),
	visionapi.OnPoll(func(t *visionapi.Task) { log.Println(t.Status) }),
)

// Or submit and walk away — the result comes to you.
ref, err := client.AnalyzeAsync(ctx, visionapi.AnalyzeParams{
	Source:     visionapi.FilePath("contract.pdf"),
	Preset:     "contract",
	WebhookURL: "https://yourapp.com/hooks/vision",
})

Results stay retrievable for 7 days; after that GetTask returns an *APIError with code result_expired (metadata survives, the payload does not).

Verifying a delivery

Deliveries are signed. Verify over the raw bytes before parsing — a re-serialized body has different bytes and will not match.

func handler(w http.ResponseWriter, r *http.Request) {
	body, err := io.ReadAll(io.LimitReader(r.Body, 10<<20))
	if err != nil {
		w.WriteHeader(http.StatusBadRequest)
		return
	}

	event, err := visionapi.VerifyWebhook(body, r.Header.Get("X-Vision-Signature"), secret, 0)
	if err != nil {
		w.WriteHeader(http.StatusBadRequest) // never parse an unverified body
		return
	}

	w.WriteHeader(http.StatusAccepted) // any 2xx is success — ack fast, work afterwards
	go process(event)                  // event.Event is "task.completed" | "task.failed"
}

VerifyWebhook rejects a bad signature, a malformed header and a timestamp more than 5 minutes old, and accepts a delivery if any v1= part matches — which is what makes a secret rotation seamless. event.ExtractionResult() decodes the payload into the same Result type as a synchronous call. Get the secret from https://app.visionapi.io/dashboard/webhooks. Failed deliveries retry at +1 m, +5 m, +15 m and +40 m, then stop.


Errors

Every API failure is an *APIError carrying the HTTP status, the stable Code, and whatever Details the endpoint attached. Branch on Code — never on the message text, which is prose and changes.

res, err := client.Analyze(ctx, params)
if err != nil {
	var apiErr *visionapi.APIError
	switch {
	case errors.As(err, &apiErr):
		switch apiErr.Code {
		case visionapi.CodeInsufficientCredits:
			// Never retried — the balance cannot change as a result of retrying.
			log.Printf("needs %d, has %d", apiErr.Required(), apiErr.Available())
		case visionapi.CodeSyncTimeout:
			task, err := client.AnalyzeAndWait(ctx, params) // re-submit on the queue
		case visionapi.CodeUnsupportedType:
			quarantine("not an image or a PDF")
		default:
			log.Printf("%s (%d) request_id=%s", apiErr.Code, apiErr.StatusCode, apiErr.RequestID)
		}
	default:
		// *UsageError, *ConnectionError, *TaskFailedError, *TaskTimeoutError
		log.Println(err)
	}
}

For a single-code check there is a shorthand:

if visionapi.IsCode(err, visionapi.CodeInsufficientCredits) { … }
HTTP Code constants
400 CodeInvalidRequest
401 CodeInvalidAPIKey, CodeUnauthorized
402 CodeInsufficientCredits — with .Required() / .Available()
403 CodeForbidden, CodeEmailNotVerified
404 CodeTaskNotFound, CodeSchemaNotFound
409 CodeConflict
410 CodeResultExpired
413 CodeFileTooLarge, CodePageLimitExceeded
415 CodeUnsupportedType
422 CodePDFEncrypted, CodeInvalidPageSelection, CodeInvalidSchema, CodeSchemaFieldConflict, CodeTooManyQuestions
429 CodeRateLimited — with .RetryAfter()
500 CodeInternalError — with .RequestID
502 CodeProviderError
504 CodeSyncTimeout

Retries and idempotency

The client retries 429, 500, 502 and network failures — three attempts by default, with the server's own Retry-After honored on 429 and exponential backoff with jitter elsewhere. Input errors and insufficient_credits are never retried, because they cannot succeed.

Every billable POST is sent with a generated Idempotency-Key, so a retried upload replays the first response instead of paying twice. Supply your own when the caller may retry — a job that re-runs, a queue that redelivers — because a fresh process generates a fresh key:

params.IdempotencyKey = fmt.Sprintf("invoice-%d", invoiceID)

Reusing a key with a different payload is a 409 CodeConflict, which is the mechanism working: it means the key already stands for something else.


Configuration

client, err := visionapi.New(
	visionapi.WithAPIKey(os.Getenv("VISION_API_KEY")), // default: $VISION_API_KEY
	visionapi.WithBaseURL("https://api.visionapi.io"), // default
	visionapi.WithHTTPClient(&http.Client{Timeout: 120 * time.Second}),
	visionapi.WithMaxRetries(3),
	visionapi.WithAutoIdempotency(true),
	visionapi.WithHeader("X-Trace-Id", traceID),
)

Every call takes a context.Context, so cancellation and deadlines work the way you expect — including during the sleep between retries and between task polls.


Account and usage

credits, _ := client.Credits(ctx)
// credits.Buckets are spent in order: Subscription → Rollover → Pack → Welcome

cursor := ""
for {
	page, err := client.Requests(ctx, 100, cursor)
	if err != nil { break }
	for _, r := range page.Data {
		fmt.Println(r.CreatedAt, r.Endpoint, r.Preset, r.CreditsUsed)
	}
	if page.NextCursor == "" { break }
	cursor = page.NextCursor
}

Usage history is metadata only — never the file, never the extracted values. Uploaded files are never retained: a synchronous request holds yours in memory for the length of the call, and an async request stages it only until the worker finishes with it.


Limits

Same for everyone:

Limit Value
Max file size 20 MB
Max PDF pages per request 50
Sync request timeout 60 s

Per plan:

Limit Free Starter Growth Pro Scale
Requests per minute, per key 10 60 120 300 600
Burst capacity 20 120 240 600 1,200
Concurrent async tasks 1 4 8 16 32
Active API keys per account 1 5 10 20 50
Saved schemas 3 10 25 100 unlimited
Max questions per Ask 5 5 5 10 10

The rate-limit bucket is per API key, not per account — splitting a workload across keys splits the limit too. The concurrency cap is per account and does not split that way: over it, an async submission answers 429 too_many_tasks and is charged nothing. Higher limits on paid plans: https://visionapi.io/pricing.


Examples

Runnable programs in examples/:

Directory What it shows
analyze/ The smallest useful call, and how to read the result
customschema/ Custom fields, line-item injection, saved schemas
detect/ Routing a mixed inbox before spending on extraction
asyncbatch/ A folder of long PDFs, queued with bounded concurrency
webhook/ A verified receiver, in one file
ask/ Visual Q&A and the Verdict field
export VISION_API_KEY=sk_live_…
go run ./examples/analyze invoice.pdf

Development

go test ./...   # offline: a httptest server stands in for the API
go vet ./...
gofmt -l .

Contributing

Issues and pull requests are welcome at https://github.com/devrobotlabs/visionapi-go. For anything about the API itself — a preset, a limit, an error code — https://support.visionapi.io reaches the team faster.

License

MIT © Vision API

About

Official Go client for the Vision API. Extract structured JSON from images and PDFs, with a confidence level on every field. No dependencies.

Topics

Resources

Contributing

Security policy

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages