SDKs

The official SDKs wrap the public API with the same surface in every language: render a template, wait for large documents, read templates and datasets, typed errors with the correlation id, safe retries and webhook signature verification.

Install#

1dotnet add package Prynt.Client

Render a document#

RenderAsync / render send the template code, the parameters and the environment, and return the document bytes with their metadata: version, page count, environment, correlation id and the run in the Logs.

1using Prynt.Client;2 3using var prynt = new PryntClient(Environment.GetEnvironmentVariable("PRYNT_API_KEY")!);4 5var pdf = await prynt.Documents.RenderAsync(6    "DDT_STANDARD",7    new { DOCUMENT_ID = 82422 },   // parameter names are sent as written8    environment: "production");9 10await pdf.SaveAsync("ddt-82422.pdf");11Console.WriteLine($"v{pdf.Version}, {pdf.PageCount} pages, correlation id {pdf.CorrelationId}");12 13// ASP.NET Core: builder.Services.AddPrynt(o => o.ApiKey = builder.Configuration["Prynt:ApiKey"]);14// then inject PryntClient.

Large documents and jobs#

When a document is too large or slow for a synchronous answer, the API replies 202 Accepted with a job. The SDKs handle it for you: they poll the job with growing intervals and download the document through its signed link, without sending your key. To manage the job yourself, start it explicitly:

1var job = await prynt.Documents.StartRenderJobAsync(new RenderRequest("DDT_STANDARD")2{3    Parameters = new { DOCUMENT_ID = 82422 },4    CorrelationId = "erp-ddt-82422",5});6 7var done = await prynt.Jobs.WaitAsync(job.JobId);         // throws PryntJobFailedException on failure8var document = await prynt.Documents.DownloadAsync(done); // renews an expired link by itself

Templates and datasets#

1foreach (var template in await prynt.Templates.ListAsync())2    Console.WriteLine($"{template.Code} v{template.Version}");3 4var rows = await prynt.Datasets.ExecuteAsync("DDT_STANDARD", "RIGHE", new { DOCUMENT_ID = 82422 }, maxRows: 100);

Errors#

Every error answer is an RFC 9457 problem; the SDKs throw it as a typed error with errorCode, detail, the HTTP status, the correlationId to give to support and, for 429, the Retry-After delay.

1try2{3    await prynt.Documents.RenderAsync("DDT_STANDARD", new { DOCUMENT_ID = 82422 });4}5catch (PryntApiException e)6{7    // e.Status, e.ErrorCode, e.Detail, e.CorrelationId, e.RetryAfter, e.Extensions8    Console.Error.WriteLine($"{e.ErrorCode}: {e.Detail} (correlation id {e.CorrelationId})");9}

Retries#

• Reads (jobs, templates, document downloads, datasets) are retried on 429, 502, 503, 504, network errors and timeouts, with exponential backoff and jitter, always honoring Retry-After.
• Renders are retried only when Prynt guarantees nothing was generated: 429 RATE_LIMITED, 503 RENDER_BUSY or DATA_SOURCE_UNAVAILABLE. After a network error or a timeout the document may already exist, so the error is returned to you instead of rendering twice. Pass your own correlation id to find the render in the Logs.

Webhooks#

Both SDKs verify the Prynt-Signature header (HMAC-SHA256 of {t}.{raw body}, constant-time comparison, 5-minute tolerance) before you parse the event.

// The raw body, before any JSON binding.if (!WebhookSignature.Verify(rawBody, request.Headers["Prynt-Signature"], secret))    return Results.Unauthorized();var evt = PryntWebhookEvent.Parse(rawBody);   // deduplicate on evt.Id

// Express: express.raw({ type: "application/json" }) keeps the raw body.if (!(await verifyWebhookSignature(req.body, req.get("Prynt-Signature"), secret))) return res.sendStatus(401);const event = parseWebhookEvent(req.body); // deduplicate on event.id