Schedule HTTP jobs with confidence

TaskPulse is a scheduler for outbound HTTP calls. You describe a request — URL, method, headers, JSON payload, and credentials — and when it should run. TaskPulse queues each execution reliably, calls your endpoint, and records the full result.

Typical uses: triggering nightly backups, pinging sync webhooks, calling report endpoints on weekdays, or re-running a failed integration on demand.

From zero to a tested job in about two minutes.

1. 1Create an account and verify your email — verification is required before jobs can be scheduled.
2. 2Optionally create a project (for example “Production webhooks”) to group related jobs.
3. 3Select New job and enter a public HTTP or HTTPS endpoint, the method, headers, and JSON payload.
4. 4Choose one-time, rate, or cron timing and confirm the timezone.
5. 5Save the job, then press Run now for a safe first test without waiting for the schedule.
6. 6Open Run history to inspect the HTTP status, response body, duration, or error message.

Projects keep jobs for different applications or environments separate. Create one project per product or environment — for example “Production webhooks” and “Staging webhooks” — and select the project while creating each job.

• All jobs remain visible under Jobs, where you can filter by project.
• A job can belong to at most one project, and you can move it at any time by editing the job.
• Deleting a project never deletes its jobs — they simply become unassigned.

TaskPulse supports GET, POST, PUT, PATCH, and DELETE. The target must be a public HTTP(S) URL — localhost and private network addresses are blocked in every environment. Headers and payloads are JSON objects; the editor validates as you type, supports Tab indentation, and can format or minify valid JSON.

Example headers

{
  "content-type": "application/json",
  "x-event-source": "taskpulse"
}

Example payload

{
  "event": "sync.requested",
  "source": "taskpulse",
  "data": {
    "mode": "incremental"
  }
}

For GET requests the payload is omitted. A content-type: application/json header is sent by default and can be overridden. Requests time out after your configured limit (30 seconds maximum), and the stored response body is capped at 2,000 characters.

Three schedule types cover one-off tasks, fixed intervals, and calendar rules. All of them run in the IANA timezone you select, so 09:00 means 09:00 in your timezone — not UTC.

UI:         Jul 20, 2026 09:30 · Asia/Kolkata
Scheduler:  at(2026-07-20T09:30:00)

UI:         Every 15 minutes
Scheduler:  rate(15 minutes)

UI:         Every 1 day
Scheduler:  rate(1 day)

0 9 ? * MON-FRI *    # Weekdays at 09:00
0 2 * * ? *          # Every day at 02:00
30 8 1 * ? *         # The 1st of each month at 08:30
0 */6 * * ? *        # Every 6 hours

Rate and cron schedules require the Pay As You Go plan; the Free plan supports one-time schedules only. See Plans and billing.

Four authentication modes cover most APIs. Credentials are encrypted with AES-256-GCM before they are stored and are never included in scheduler or queue messages. Editing a job keeps its existing credentials unless you explicitly enter new ones.

POST /api/backup HTTP/1.1
Host: example.com
Authorization: Bearer sk_live_9f8e7d6c5b4a
Content-Type: application/json

POST /api/report HTTP/1.1
Host: example.com
X-API-Key: 4f2a09c1-7b3e-4d55-9a10-2f8f4be6d001
Content-Type: application/json

X-TaskPulse-Timestamp: 2026-07-15T09:30:00.000Z
X-TaskPulse-Signature: sha256=88a1c0e6…

import crypto from "node:crypto";

function verifyTaskPulse(req, secret) {
  const timestamp = req.headers["x-taskpulse-timestamp"];
  const received = req.headers["x-taskpulse-signature"]; // "sha256=<hex>"
  const expected = "sha256=" + crypto
    .createHmac("sha256", secret)
    .update(timestamp + "." + req.rawBody)
    .digest("hex");
  // Reject stale timestamps (e.g. older than 5 minutes) to stop replays.
  const fresh = Date.now() - Date.parse(timestamp) < 5 * 60 * 1000;
  return fresh && crypto.timingSafeEqual(
    Buffer.from(received),
    Buffer.from(expected),
  );
}

Every job keeps its schedule, manual controls, and history together, so day-two operations never require digging through infrastructure.

Retries: when a run fails (non-2xx response, timeout, or network error), delivery is retried automatically up to your configured maximum attempts. Messages that keep failing are moved to a dead-letter queue instead of retrying forever.

Each execution creates a run record you can open from the job page (5 runs per page) or the dashboard. A run moves through queued → running → success or failed.

TaskPulse has one free tier and one usage-based tier — no subscriptions, no per-seat pricing.

How credits work

1. 1Open Account → Usage and billing and pick a credit pack (or a custom multiple of 100).
2. 2Pay with Stripe (card) or Razorpay (card, UPI, netbanking, wallet). Your first purchase activates Pay As You Go automatically.
3. 3Credits are added the moment payment is confirmed; your remaining balance, usage, and dollar value stay visible on the Account page.
4. 4When the balance reaches zero, new runs are blocked (existing configuration is kept) until you top up again.

100 invocations   = $1
500 invocations   = $5
1,000 invocations = $10
Every recorded run (including each retry attempt) = 1 invocation

TaskPulse is designed so that a leaked message queue or log line never exposes your credentials.

• Auth secrets are encrypted (AES-256-GCM) at rest and excluded from scheduler and queue messages — messages carry only the job id.
• Only public http(s) URLs are allowed; localhost and private/internal network ranges are rejected at creation time and again at execution time.
• Requests are limited to a 30-second timeout, and stored response bodies are truncated to 2,000 characters.
• Sessions are server-managed HTTP-only cookies; email verification is required before jobs can be created or run.
• Payment card data never touches TaskPulse — checkout happens on Stripe or Razorpay, and Razorpay callbacks are signature-verified server-side.

Start with the run record — it almost always names the problem. These are the most common situations:

Still stuck? The FAQ covers pricing and payment questions, and every page of the app links back to this guide.