Postlad/Guides/Replacing a platform that shut down

Replacing an IoT platform that shut down

If you are here because a device has been posting into nothing, start with the export rather than with the choice of replacement. Getting the history out is the reversible step, and it is the one that stops mattering the day the domain lapses. Everything below assumes you will be moving again some day — including away from us.

Guide·8 min read·Last checked August 2026

First, get the data out

Do this before you evaluate anything, and do it even if you end up somewhere other than here.

1. Open the old dashboard and export every channel or feed you care about, in whatever format it offers. A CSV with timestamps is worth more than a tidy one without.
2. Keep the download untouched as your archive copy. Work on duplicates.
3. Write down what the columns mean — units, sensor positions, which field was which. A file of numbers labelled field1 is worth much less in a year.
4. Screenshot the dashboard settings you would otherwise have to reconstruct from memory.

If the service is already unreachable, there is usually nothing to recover, and we are not going to pretend otherwise. Check whether an old browser tab, a phone app cache or a local logger still holds a copy before you accept that.

What we found, and when

On 6 August 2026 we ran DNS and HTTP probes against the hobby platforms that comparison sites still list. This is what came back. Where we could not establish a date, the table says so instead of guessing.

Two of those rows are worth reading together. Phant's shutdown notice pointed its users at three replacements. Of those three, one is the Blynk platform whose legacy servers were switched off at the end of 2022, and one is Cayenne, which no longer has a consumer front door. This cohort was not orphaned once. It was routed onward and orphaned again.

That is the part we take seriously, and it is why the rest of this page is mostly about exports and guarantees rather than about features.

Not dead, but smaller

Some of what people are searching for is not a shutdown at all — it is a free tier that no longer fits the project. From the same 6 August 2026 check of vendors' own pricing pages:

• Arduino Cloud free: 2 Things, 1 day of data retention, 25 sketch compiles a day.
• Blynk free: 5 devices, 1 user, 1 week of retention, 100,000 messages a month, and a dashboard that shows a single device.
• Ubidots: no free plan listed on the pricing page; the entry plan was $99 a month.
• InfluxDB Cloud free: 2 buckets and 30 days.

A greenhouse logged over a season does not fit in a day, a week or a fortnight. That gap is the reason this product exists, and the pricing page states our own numbers in the same form so you can hold them against us later.

Why this shape probably fits what you lost

The platforms in that table had something in common that most modern ones do not: a device could write to them with one plain URL. If your sender is a few lines of firmware that build a query string, the services that replaced them mostly ask you to adopt an SDK, a broker or a line protocol first.

Postlad is deliberately the old shape:

• One HTTP GET writes a reading. http://api.postlad.com/u?k=KEY&f1=23.5 — no SDK, no auth header, no broker.
• Plain HTTP is a first-class route, not a fallback, because an ESP8266 with a certificate bundle is a different project from an ESP8266 with a sensor on it.
• A public link per stream. Send someone the URL; they need no account to look at the chart.
• No device management, no app builder, no rules engine. The product is: it arrives, it is drawn, it is kept, you can take it away.
• Honest refusals. Nothing returns success for a reading it did not store. A rejected write says what was wrong in one sentence and what to do about it.

If your old sender was a ThingSpeak-style URL, there is a shorter road: point your ThingSpeak sketch at Postlad — the /update path accepts api_key and field1–field16 directly.

What we cannot promise, and what is written down instead

We cannot promise longevity. Postlad launched in 2026; anyone claiming a track record after a few months is selling one. What we can do is make leaving cheap, and put that in writing where it can be held against us:

• CSV export works on every plan, at any time — including from a stream that has gone read-only, and after a downgrade. The terms say it is deliberately not a setting we can change our minds about.
• CSV import works on every plan too, so the door swings both ways and your history is not stranded by the move.
• Retention is stated as a number, not as "unlimited". On Free, every reading is kept at full detail for 30 days. After that it becomes an hourly minimum, maximum and average, kept for 12 months. Older than that is deleted.
• Deleting a stream is reversible for seven days, and the error you get from a device writing to a deleted stream tells you that, with the date.
• A status page that is not hosted on the thing it reports on: postlad.com/status.
• Your data is yours. It is not sold and it is not used to train anything.

None of that makes us permanent. It makes the cost of us being wrong an afternoon's work rather than a lost season.

Moving a sender that has nowhere to post

There are two shapes of migration and they are not the same amount of work. Say which one you have before you estimate the job.

If your device builds a URL — a query string with a key and one or more numeric parameters — you are changing a string. Point it at http://api.postlad.com/u, rename the key parameter to k and the value parameters to f1–f16, and read the reply. That is the whole change, and the reply will tell you precisely which parameter it did not recognise if you miss one.

GET http://api.postlad.com/u?k=pk_live_YOUR_KEY&f1=23.5&f2=61
→ HTTP 200
  ok 1

If your device used a vendor library — Blynk's virtual pins are the common case — then there is no URL to repoint, and we are not going to describe that as a two-minute job. The send path gets rewritten as one HTTP request; on an ESP32 that is roughly ten lines, and the ESP32 guide has them. What you gain is a sender with no library to go unmaintained under you, which, on this page in particular, is not a small thing.

Either way, the order that avoids surprises:

1. Create the account and the stream, and copy the write key.
2. Send one reading by hand — from a browser, curl, or the board — and confirm the chart moves.
3. Import the archive CSV, and check the first and last timestamps against the file.
4. Change the device, and leave the old destination configured until you are satisfied.
5. Create a stream-goes-quiet alert, so the next outage arrives as an email rather than as a gap you find later. On Free you get two rules and email delivery; see store and forward for the version where the device also keeps its readings through the outage.

How to judge the next one

Apply this to us as readily as to anyone else. A service that cannot answer these in public is a service you should assume you will be exporting from in a hurry:

• Are the free-tier limits written down as numbers, on a page that loads?
• Does it say how long data is kept — and what happens after that?
• Can you export everything, without asking, today?
• Can you get data back in, so a migration is not one-way?
• Is there a status page, and does it live somewhere else?
• When a write is refused, does the service say so, or does it answer success and drop the reading?
• Is the price of the next tier up visible without a sales conversation?

We would rather be judged on those than on a feature grid, because the graveyard above is not full of platforms that lacked features.

FAQ

What is a dweet.io alternative in 2026?

Anything you pick, pick it knowing that dweet.io's domain was parked when we probed it on 6 August 2026 and that recommendation lists have not all caught up — one prominent comparison site still named it the best ThingSpeak alternative at that date. If what you liked was posting a reading with a single URL and getting a link back, that is the shape of this product; the write is http://api.postlad.com/u?k=KEY&f1=23.5 and the stream page is public if you make it public.

What replaced SparkFun's Phant / data.sparkfun.com?

Nothing official that is still both free and running. data.sparkfun.com had no DNS record at all when we checked on 6 August 2026, and the three services Phant's own notice recommended have not all held up — see the table above. Phant's own code is still published, so self-hosting is a real option if you want to run the server yourself. If you would rather not, a hosted logger you write to with one URL is the closest match to what the service did.

Can I still run a Blynk legacy server?

Community forks of the discontinued legacy server exist, and their own descriptions say plainly what they are forks of. Running one is a real option if you want a local server and are willing to own it. If the appeal was that somebody else ran it, that is the part the forks cannot restore, and a hosted logger with an export button is the honest replacement.

Is my data recoverable from a service that has already gone?

Usually not, and we are sorry — that answer is the reason this page leads with the export. Look for an old export in your downloads folder, a phone app that cached recent readings, or a local logger that was writing in parallel. If one turns up, it can be imported here with its timestamps intact.

Why should I believe this one lasts?

Do not believe it. Check the export button works on your first day, keep a copy of your data somewhere you control, and hold us to the numbers on the pricing page. That is the right posture towards every service on this page, ours included.

Start with one reading

An email address and a magic link, then a write key. Send one number before you migrate anything — the whole point of this shape of product is that the first test takes a minute.

Create your free account →

Related guides

• Point your ThingSpeak sketch at Postlad: the /update alias, and every divergence we know of.
• Move a CSV history across: preview, field mapping and what happens to old readings.
• Store and forward after an outage: keep the readings a network failure would otherwise cost you.
• Send ESP32 data to the cloud: the ten-line sender, if you are replacing a library.
• Raspberry Pi data logger: a logger with a durable local spool, for sites where the network is the unreliable part.

Sources and verification

• Platform statuses on this page come from direct DNS and HTTP probes run on 6 August 2026, plus the vendors' and projects' own published pages. They describe that date and are not re-probed continuously; if you find a row that has changed, hello@postlad.com reaches a person and the row gets fixed.
• Where a shutdown date could not be established from a primary source, the table says so rather than repeating a date from a comparison site.
• Phant discontinuation notice: https://github.com/sparkfun/phant
• Free-tier figures for Arduino Cloud, Blynk, Ubidots and InfluxDB Cloud were read from those vendors' own pricing pages on the same date.
• Postlad's own commitments on this page are the ones written on /terms and /pricing, which are asserted against the running product by our test suite.