tracebin
Support artifact collection

Collect diagnostic files from customer machines without the back-and-forth.

Customer terminalsh
$ curl -sSL https://tracebin.dev/c/<token>/run | sh

Collected files:
  [1] var_log_app_app.log            2.1 MB  (host:/var/log/app/app.log)
  [2] etc_myapp_config.yml           4.2 KB  (host:/etc/myapp/config.yml)
  [3] var_log_nginx_error.log      812.4 KB  (container:api:/var/log/nginx/error.log)

Total: 3 files, 2.9 MB

Upload 3 files (2.9 MB)? [y/N/review]
Start freeFree during beta — no payment details required.
The workflow

Getting files out of a customer environment

Without tracebin
  • An email thread asking the customer to go find /var/log, zip it, and send it back.
  • The attachment bounces at the mail server's size limit, so it moves to a personal file-sharing link that expires.
  • A zip arrives unlabeled — wrong host, wrong day, no hashes, no way to tell what it actually contains.
  • Every missing file is another round trip, and every round trip is another day.
With tracebin
  1. 1Send one link.
  2. 2The customer runs one command.
  3. 3They review exactly what was collected and approve the upload.
  4. 4Files arrive in your console, organized per run — names, sizes, hashes, exit codes.
How it works

Four steps, no software installed

01

Create a bin

Define a collection spec: file globs, commands to capture, optionally a Docker container to collect from.

02

Send the link

tracebin generates a one-liner tied to that single bin. Tokens are unguessable and expire on a schedule you set.

03

Customer approves at the manifest gate

Before anything uploads, the script lists every collected file with its size and source. The customer can exclude items or decline — no approval, no upload.

04

Runs appear in the console

Each run stores its manifest: file names, sizes, SHA-256 hashes, source paths, command exit codes, timestamps.

Container collection

Globs expand and commands run inside a named Docker container, with exit codes captured per command.

Wrap mode

The link can wrap a reproduction command: the customer consents up front, the command runs, and collection starts when it exits — including the command's output tail.

Trust

Built to be run on machines you don't own

Approval before upload

The customer sees every file — name, size, source — and nothing uploads until they approve.

Inspectable script

Plain shell, no binaries. The customer can download the script and read it before running it instead of piping to sh.

Direct-to-storage uploads

Files go straight to object storage over time-limited signed URLs. They don't pass through anyone's inbox.

Expiring links

Collection tokens are unguessable, scoped to a single bin, and expire. Expiry is enforced on the server, not in the script.

No customer account

The person running the collection never signs up or logs in. The link is all they need.

More detail

A dedicated security page is coming. Until then, questions go to hello@tracebin.dev.

Pricing

Free while in beta

The full product is free during the beta — no payment details, no feature gates.

Free — during beta
$0during beta
  • Full product — collection specs, containers, wrap mode
  • 5 GB storage
  • 30-day retention
Start free
Team — planned
$20/user/mo — indicative
  • Teams with shared bin visibility
  • Pooled storage
  • Bring your own S3/R2 bucket
  • 1-year retention

In development. Pricing is indicative — there is no billing yet.

Contact us