Hosted background-job platform — 7 official SDKs

Durable background jobs for every stack, without the infra.

Enqueue jobs from any service — official SDKs for .NET, TypeScript, Python, Go, Java, PHP, and Ruby — and let Zeridion handle the queue, retries, recurring schedules, and dead letters. Your workers pull and execute. No queue infrastructure to babysit.

If your jobs run on an in-process timer or a hand-rolled retry loop, and you’d rather ship features than babysit them, this is for you.

Start free →Read the quick start
Live status →Already a customer? Sign in
Program.cs
// Install the SDK
dotnet add package Zeridion.Flare --prerelease

// Wire it up in two lines
builder.Services.AddZeridionFlare(opt => {
    opt.ApiKey = builder.Configuration["Zeridion:ApiKey"];
    opt.ApiBaseUrl = "https://api.zeridion.com/flare/v1";
});

app.UseZeridionFlare();
Official SDKs
.NETTypeScriptPythonGoJavaPHPRuby
Platform

Why teams choose Flare

Designed for operational clarity from the first queued job to production scale.

One managed endpoint, zero queue infrastructure

Drop the SDK in and make one HTTPS call to api.zeridion.com/flare/v1; storage, scheduling, retries, and the dashboard are managed.

Feels like the .NET you already write

IJob<T> classes with constructor DI, async/await, and CancellationToken. One line to enqueue, strongly typed payloads. No expression trees, no magic strings — just an optional [JobConfig] attribute when you want per-class defaults.

Production defaults, not homework

Exponential backoff with jitter, dead-letter on exhaustion, a stuck-job reaper, and an Idempotency-Key header that replays the original response on duplicates — all on by default.

Chain jobs without an orchestrator

Mark a child to run only after its parent succeeds with ContinueWithAsync. Failures and dead-lettering cascade automatically. No Durable Functions, no Temporal, no DIY chaining through a second queue.

Observability that joins your traces

Every job emits OpenTelemetry spans tagged with tenant.id and tenant.plan. A metrics API exposes summary, throughput, and queue depth — drop straight into Grafana, DataDog, or Honeycomb.

Backoff that knows when to retry

Per-tenant sliding-window rate limits enforced atomically in Redis. A 429 surfaces as FlareRateLimitException with ResetAt — your client waits the exact window, not a guessed exponential.

How it works

Four steps from one line of code to a dashboard row.

You write the enqueue call and the job class. Flare owns the durable store, the pull loop, retries with backoff, and the dead-letter path — and reflects every state change in the dashboard.

  1. Enqueue from your code

    One line via the NuGet SDK, with a strongly-typed payload. No queue client, no broker connection string.

    await jobs.EnqueueAsync<SendInvoice>(payload);

  2. We store it durably

    The Flare API persists the job to a managed datastore. You ship no queue infrastructure, no broker, and no DB migration of your own.

  3. Your worker pulls and runs

    Your process picks up the next eligible job and executes it. On failure we schedule a retry with exponential backoff + jitter; after MaxAttempts the job moves to dead-letter — automatically.

  4. Observe everything

    Every state change lands in the hosted dashboard. OpenTelemetry spans tagged with tenant.id flow into your tracing stack, and the metrics API exposes throughput and queue depth.

5-minute quickstart

From dotnet add to first job

Copy each step. Run against the live API. No leaving this page.

dotnet add package Zeridion.Flare --prerelease

// appsettings.json
{
  "Zeridion": {
    "ApiKey": "zf_live_sk_xxxxxxxxxxxxxxxxxxxx"
  }
}

Automatic retries

Exponential backoff with configurable retry counts and dead-letter handling.

Progress reporting

Call ctx.ReportProgress(0.42) from your job; the value rides the next heartbeat to the dashboard’s progress bar.

Recurring jobs

Define cron schedules directly in code — no external scheduler or timer service required.

Queue visibility

Dashboard metrics, alerting, and full payload inspection for every job state.

Pricing

Simple, predictable pricing

Start free. Upgrade when your jobs outgrow the tier you’re on.

Free

$0

forever

  • 50,000 jobs / month
  • 1,000 requests / hour
  • 1 project
  • Hosted dashboard
  • Community support
Start free

Starter

$29

/ month

  • 300,000 jobs / month
  • 5,000 requests / hour
  • 3 projects
  • Opt-in overage: $1.50 / 10k extra jobs
  • Email alerts
Choose Starter
Most popular

Pro

$99

/ month

  • 3,000,000 jobs / month
  • 50,000 requests / hour
  • Unlimited projects
  • Opt-in overage: $1.00 / 10k extra jobs
  • Priority support
Choose Pro

Business

$299

/ month

  • 30,000,000 jobs / month
  • 500,000 requests / hour
  • Unlimited projects
  • Opt-in overage: $0.50 / 10k extra jobs
  • Custom data retention
Choose Business

Enterprise

Custom

  • Custom job volume
  • Custom rate limits
  • Unlimited projects
  • SLA & SOC 2 report
  • Dedicated support
Talk to sales

Allowances are monthly. Paid plans can enable opt-in overage to keep running past the allowance — with an optional spend cap so a bill never surprises you. Without overage, new enqueues return HTTP 429 until the allowance resets at the end of your billing period (calendar month on Free).

Need a custom contract, dedicated support, or a security review?

We work directly with enterprise teams on signed DPAs, security questionnaires, and named contacts for support and onboarding. Audit-log export ships on every plan today; SSO/SAML is on the roadmap below — talk to us if it is a hard requirement and we’ll sequence the work against your timeline.

Talk to sales →

On the roadmap for paid plans: SSO/SAML, Slack alerts, priority queues. (Audit-log export shipped in 0.1.2 — available on every plan today.)

FAQ

Common questions

Quick answers before you talk to anyone. If something here isn’t clear, the docs go deeper.

Do I have to host workers myself?
You host the job code, we host everything else. In .NET, the Zeridion.Flare NuGet package registers a hosted service inside your existing ASP.NET Core app that polls our API for work — no separate worker process needed. In TypeScript, Python, Go, Java, PHP, and Ruby, the SDK is a REST client you call from your existing service or a small worker process. The data plane (queue, scheduler, retries, dashboard) is fully managed in all cases.
Which .NET runtime versions are supported?
The SDK multi-targets net10.0 and netstandard2.1, so it runs on any project that supports .NET Standard 2.1 or later — including .NET 6, 7, 8, 9, and 10 on Windows, Linux, macOS, and ARM. ASP.NET Core is recommended but not required.
How is this different from running a job library in-process?
An in-process library couples job durability to your web tier’s deployment, scale, and storage. Flare gives you a managed queue and dashboard that survive deploys and process restarts, exponential-backoff retries with dead-lettering on by default, and a real-time UI for inspecting payloads and failures — without standing up Redis, Postgres, or a worker host of your own.
What happens when I hit my plan’s monthly allowance?
New POST requests to create jobs return HTTP 429 with error code monthly_allowance_exceeded and X-Quota-* response headers showing your allowance, current usage, and the reset time — the end of your billing period on paid plans, calendar month-end UTC on Free. Jobs already in the queue continue to run. Paid plans can enable opt-in overage to keep creating jobs at the plan’s per-10k rate, with an optional spend cap; once the cap is reached, creates return spend_cap_reached until you raise it or the period resets.
Is there a free tier, and does it stay free?
Yes — 50,000 jobs per month, one project, hosted dashboard, $0/month, no credit card. The free tier is part of the published pricing model, not a trial; it doesn’t expire and doesn’t auto-convert.
Can I run Flare offline or on-premises?
Not today. Flare is a hosted service; the API and dashboard run on our infrastructure. If a fully self-hosted deployment is a hard procurement requirement, tell us — we want to hear the use case, but we can’t promise it on the roadmap.
How is my data stored — what about GDPR?
Job payloads and metadata are stored in managed Postgres on Azure with TLS in transit and encryption at rest — the queue itself is Postgres-backed, so payloads never leave it. Redis holds only transient rate-limit counters and short-lived cache entries. We can sign a DPA referencing the EU SCCs — request one here. Region pinning beyond the default is in progress.
What if I want to leave?
Every SDK is a thin client over a fully documented REST API, so you’re never blocked on us to read your data or wire a different client. Job payloads are your JSON — export from the dashboard or the API. Cancel any time; the free tier just keeps running.
Get started

Ship background work in minutes, not sprints.

Install the NuGet package, add two lines of config, and start enqueuing jobs against the hosted Flare API.

Create your account →Read the quick start