A document you can shape
Any JSON object: nested settings, arrays of hosts, readings or counters. No schema, no migrations — the record is whatever your software needs it to be.
One JSON record per account — with a backup
webrecord keeps exactly one JSON object per account: device configuration, last-known state, a small inventory or the snapshot your script needs tomorrow. Send the whole document in one request, merge a single fragment at a dotted path, and always find the previous version waiting in the backup — without running a database of your own.
The idea
Most small projects do not need a database cluster. They need one dependable place to keep a JSON document — and to read it back later from a different process, network or continent.
Any JSON object: nested settings, arrays of hosts, readings or counters. No schema, no migrations — the record is whatever your software needs it to be.
Every save or update moves the previous document into the backup slot. One request to
/webrecord/get-backup and yesterday's version is back.
Change one setting with a dotted path instead of resending megabytes. Objects merge,
scalars replace, null removes the key.
How it works
Registration is e-mail based, because the token is the whole account.
01
Send your address to /webrecord/register. We reply with a
token, a ban-unlock token and a confirmation link.
02
Open the link. From that moment the token can write and read — until then it is inert.
03
/webrecord/save replaces the document,
/webrecord/update merges one fragment, and
/webrecord/get-backup returns the previous version.
curl -X POST https://webrecord.analogbrain.com/webrecord/save \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" \
-d '{"record":{"device":"sensor-7","counter":1842,"online":true}}'
curl -X POST https://webrecord.analogbrain.com/webrecord/update \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" -d '{"path":"device.online","value":false}'
curl -X POST https://webrecord.analogbrain.com/webrecord/get-backup \
-H "Authorization: Bearer $TOKEN" \
-H "Content-Type: application/json" -d '{}'
{"record":{"device":"sensor-7","counter":1842,"online":true},
"timestamp":"2026-01-01T12:00:00.000Z","bytes":61}
API
Send the token in the Authorization: Bearer header. No SDKs, no
versioning rituals — JSON in, JSON out.
| Method | Path | Auth | What it does |
|---|---|---|---|
| POST | /webrecord/register | — | Start registration, tokens delivered by e-mail |
| GET | /webrecord/confirm | token in link | Confirm the e-mail address and activate the token |
| POST | /webrecord/save | Bearer token | Replace the whole JSON record, keep the old one as backup |
| POST | /webrecord/get | Bearer token | The current record, its UTC timestamp and byte size |
| POST | /webrecord/get-backup | Bearer token | The previous version of the record |
| POST | /webrecord/update | Bearer token | Merge a fragment at a dotted path (deep merge, null removes) |
| POST | /webrecord/unban | Bearer + unban token | Lift an IP ban for the calling address (20 uses) |
| POST | /webrecord/delete | Bearer token | Delete the account, its record and its backup |
| GET | /healthz | — | Service and database health |
Machine-readable schema: /webrecord/openapi.json
Limits & fair use
Limits exist so one misbehaving client cannot degrade the service for everyone else.
One JSON object per account: 512 KB on Free, 1 MB on Pro, 5 MB on Business. The limit is
counted in UTF-8 bytes of the compact JSON; an oversized write is rejected with
413 and changes nothing.
Each account keeps the current record plus the previous version. A save or update rotates the backup; reading it never consumes anything.
10 requests with a bad token ban the IP for 15 minutes, then 24 hours, then 7 days. More than 5 registrations per second bans for 15 minutes; 15 unknown paths ban for 24 hours.
Each account carries a separate unban token (20 uses) that lifts the ban for its current IP — spent only when a ban is actually removed.
Pricing
Every plan stores the current record and one backup copy, timestamped in UTC. Switching plans changes the limit of an existing account — no migration, no new endpoint.
$0/month
One JSON record and its backup for scripts, sensors and side projects.
Start free$5/month
A full megabyte for configuration documents and device inventories.
Request Pro$7/month
Five megabytes for fleets, snapshots and data-heavy pipelines.
Request BusinessPaid plans are activated on your account after a short e-mail exchange — the API stays the same.
FAQ
Short answers about the free tier, limits and how an account is created.
No. webrecord stores one JSON record per account for you: your app sends the whole document with one HTTPS POST and reads it back whenever it needs it. No schema, no migrations, no backup job.
The free plan allows a single record of up to 512 KB. Pro allows 1 MB and Business 5 MB.
The size is the UTF-8 length of the compact JSON, and an oversized write is rejected with
413 without touching the stored document.
Each account keeps two copies: the current record and the version that was current before
the last successful save or update. /webrecord/get-backup returns that
previous document with its timestamp; a save whose content is unchanged does not rotate it.
/webrecord/update takes a dotted path and a value.
Objects are merged deeply, scalars replace the old value, array elements are addressed by
index and null removes the key. The rest of the document stays untouched.
Register an e-mail address, confirm it from the welcome message, and the API token plus a ban unlock token arrive in the same e-mail together with a markdown instruction file.
/webrecord/register, /webrecord/save,
/webrecord/get, /webrecord/get-backup,
/webrecord/update, /webrecord/delete and
/webrecord/unban — every call is a single HTTPS POST authenticated with a
bearer token. There is no history endpoint.
Register an address, confirm it from the welcome e-mail and your account is ready — tokens, limits, copy-ready commands and a backup slot arrive with the message.