Your first pipeline

Sign in, describe the data you need, review Jason's proposal, watch the build, check the first rows, and find your way around the Home page.

This page walks through the whole first-time flow, from the sign-in dialog to a workspace with a running pipeline. Nothing here needs preparation beyond an email address and an idea of the data you want.

You need
An email address
Rows used
Up to your free 100
Time
Minutes to propose; up to an hour or two to build
Result
A workspace with a verified first dataset
From brief to live pipelineDescribe the goal, review Jason's proposal, let Jason build, check the first run, then the pipeline runs on its schedule. You can refine the brief after seeing the first rows.Describegoal + filesProposalsources, fields, scheduleBuildJason explores and buildsFirst runa sample you reviewLivescheduled + alertingContinuerefine the brief
From a brief to a verified dataset. You review at two points: the proposal and the first results.

1. Sign in

Go to factory.jsonify.com/setup. You can describe your goal before you sign in; signing in is only required when you press Build my workspace, and your brief and attachments stay saved through it.

The sign-in dialog is titled Sign in or create an account. There is one flow for both: if the address is new, an account is created for you.

  1. Enter your email address and press Continue with email. A quick security check may appear; complete it and continue.
  2. Enter the six digits from the email you receive. The number expires in 10 minutes. A resend link appears after a 30-second wait, and Change email takes you back a step.
  3. Choose your organisation if you belong to more than one. The one you used last is marked Last used.
  4. Add your name the first time. First name is required; Last name is optional. If a personal organisation was created for you, you can rename it here (the default is your first name followed by “‘s Organisation”).

Prefer Google or single sign-on? Below the email field, Google and Other options open the hosted secure sign-in page. Organisations that require SSO or multi-factor authentication are sent there automatically with a short explanation.

Sign-in uses a one-time number sent to your inbox. There is no password to set or remember.

2. Describe what you want

The setup dialog asks What would you like to collect? Type your goal into Your goal (up to 2,000 characters). Say what to collect, from where, what a row is, and how often it should refresh. You do not need to explain how any website works; Jason works that out.

The setup dialog with a brief typed into the Your goal field and an attached spreadsheet
The setup dialog. The brief can be a sentence or a paragraph; attachments are read as part of it.

Two briefs of different shapes, both of which work:

Track the price, promotion text and stock status of the 120 products in the attached sheet across Tesco, Sainsbury’s and Ocado. One row per product per retailer per day. Refresh every morning at 6am UK time and flag any price change over 5%.

Find every planning application submitted to Manchester City Council in the last 30 days. For each one keep the reference, address, proposal description, applicant, status and decision date. I don’t have a list of sources; the council portal is the obvious one but include anything else that publishes them. Weekly is fine.

The first brief is a closed list with a schedule and a rule; the second is open-ended and asks Jason to propose sources. Either way, Jason comes back with a proposal you can adjust.

Attach files opens a picker with three tabs: File (drop or browse, any file type), URL (paste a Google Sheet or web page address) and Text (paste a list, one item per line). Attachments show as green chips under the goal. Limits are 20 files per message, 25 MB each and 100 MB in total.

Advanced reveals two options. There are only two:

Option Values Default What it does
Use synthetic data on / off off Builds a dashboard from fictional sample rows instead of collecting from real sources. Jason still writes a reviewable brief, but no pipeline is created. Useful for trying the analytics side before committing sources.
Skip, just make an empty workspace link Creates a workspace named “New Workspace” with nothing in it and takes you to its Home page. Use it when you want to upload a dataset or start from a chat instead of a build.
The setup dialog with the Advanced section open, showing Use synthetic data and Skip, just make an empty workspace
The two Advanced options.

Press Continue (or Enter). While Jason reads the brief the dialog shows Understanding your goal…, then Preparing your setup…. If a goal is outside what Jsonify collects, the dialog explains why and you can rewrite it; Start a new setup clears everything.

Open factory.jsonify.com/setup later and you are offered Resume setup or Start fresh. Starting fresh keeps the earlier conversation saved.

3. Review the proposal

After the dialog closes you are in the setup conversation. Jason’s activity appears as lines you can expand, and a floating Jason View card shows the page Jason is looking at as it works. Drag the card by its title bar, enlarge it, or minimise it to a pill.

On the right, Your workspace summarises the proposal as it takes shape:

  • Sources. URL pills, up to 10. Remove one, or Add a public http or https address and press Apply sources to have Jason re-inspect.
  • Columns. Up to 30 chips with their types. Add or remove and press Apply columns.
  • Estimated Volume. A range of rows per month with a How this was estimated disclosure.
  • Schedule. The proposed cadence, marked “proposed, not yet enabled”. Edit it here or leave it for later.
  • Row pricing. Your starting allowance and the estimated monthly cost at the proposed volume, with a How does pricing work? slider.

You can also just reply in the composer. Anything you type goes to Jason as a follow-up, and Jason may ask questions of its own: a card headed Jason has a question with choices to pick from, or a box to write your own answer.

Keep the retailer list but drop the stock status column, and add the unit price per 100g where the site shows it.

When the brief is ready a blue Workspace proposal panel appears with Review proposal. The proposal opens as a dialog titled You’re ready to build! and lists:

  • the brief name and purpose,
  • Sources,
  • Reading data, one chip per field with its type and whether it is optional,
  • Estimated data volume, with an Explain disclosure,
  • Product and plan, showing Jsonify Radar, the Free plan and your monthly row quota.

Tick I accept the Terms, Privacy Policy and Acceptable Use Policy., then press Build my workspace. Revise with Jason closes the dialog and puts the cursor back in the composer instead. If you have unsent text or files in the composer, send them first so they are included.

Collection stops automatically when the monthly quota is exceeded. You can raise the limit at any time in Billing and usage.

4. Watch the build

The build runs in the background and can take an hour or two, because it ends with a real first run against your sources. A dialog headed Jason is building your workspace stays on screen with the workspace name and a note that you will be emailed when it is ready. You can close the page.

Latest Jason activity at the bottom of the dialog expands into a live log of what Jason is doing: opening pages, working out layouts, testing on a sample, and sending the candidate for independent review. The same activity is available afterwards in the build chat under Chat History.

Two things can interrupt a build:

  • A question. If Jason needs a decision, the question card appears in the dialog and the build waits for your answer.
  • A blocker. The dialog changes to Your build needs attention with Resolve with Jason, which opens the conversation so you can read what happened and reply.

5. Your workspace is ready

When the first run has been verified the dialog reads Your workspace is ready, shows the number of verified rows, and offers View data and View production run. Go to workspace takes you to Home.

Before you look at anything else, open the dataset and read the rows:

  1. Check the columns match the brief. Every field you asked for should be there with the right type. Numbers should be numbers, dates should be dates.
  2. Open a few rows. Click a row, then Origin, to see the page each value came from and a screenshot of it as collected. See Row detail and provenance.
  3. Read the statement of work. Open the pipeline. The Statement of Work says what is in scope and, just as important, what is not. See Statement of work and steps.
  4. Look at the run. Runs on the pipeline page shows pages fetched, rows collected and duration. See Run detail.

The first run is real data, not a preview. It counts towards your monthly rows, and it is the first dataset version.

6. Refine the brief

Anything that is not right is a message to Jason, either from the pipeline page (Edit pipeline) or from a new chat with the pipeline mentioned. Jason makes the change as a new pipeline version, verifies it, and activates it. The full flow is on Changing a pipeline.

The promotion text column is picking up the “Clubcard price” badge as well as the offer. Keep only the offer wording, and add a separate true/false column for whether a Clubcard price applies.

Turn the schedule on: every day at 06:00 Europe/London.

The build leaves the schedule off until you have seen the first results. Turn it on by asking Jason, or in the pipeline’s Settings. See Scheduling.

The Home page

Home is the overview of one workspace. It refreshes itself every few seconds, so a running pipeline or a working chat updates without a reload. Until the workspace has a pipeline or dataset, Home is a single composer headed What do you want to build?; once populated it shows these sections in order.

Suggested tasks

Jason proposes next steps grounded in your goal, your pipelines and their recent runs. Two cards show at a time; Show more suggestions reveals the rest and Refresh asks again. Click a card to read the detail, then Send to Jason puts the task into the Jason panel for you to edit and send (nothing is sent automatically) or No thanks dismisses it for everyone in the workspace.

The Suggested tasks section on Home with two suggestion cards and a Refresh link
Suggested tasks. Dismissals are shared across the workspace.

Needs attention

Shown only when there is something to act on:

  • Active chats. Chats where Jason is working, queued, or monitoring a run, with the elapsed time. Click to open.
  • Pipeline warnings. Pipelines whose last run failed or whose health is degraded, with the reason and how long ago. Click to open the pipeline; the X dismisses the warning until the next failure.
  • Invitations from the Jsonify team. A shared Slack Connect channel for support, and community offers. Each opens a short dialog; the X hides it for the workspace.

Pipelines

Every pipeline as a status card: name, cadence, the number of rows in its outputs, a status line and when it last ran. Card colour follows health: white before the first run, green when healthy, blue while running, amber for a warning, red after a failure, and purple when an automatic repair fixed the last failure. Next run shows the scheduled time. View all opens the pipelines list.

The Pipelines section on Home showing six pipeline status cards with row counts and health colours
Pipeline cards. A running card shows a live clock.

Status lines you will see: Not run yet, Running, Queued, Finished successfully, Repaired, Synced, Awaiting publication, Needs attention, Last run cancelled and No recent successful run.

Scheduled

The next five things due to happen: scheduled pipeline runs and any follow-ups Jason has scheduled in a chat (tagged Jason task). Each shows the time and a live countdown. Empty text: No scheduled pipeline runs or Jason tasks.

The Scheduled section on Home listing upcoming pipeline runs with countdowns
Scheduled runs and Jason tasks, soonest first.

Datasets and Deliverables

One card in two halves. Datasets lists the five most recently updated datasets with row counts and the producing pipeline. Deliverables lists files a pipeline produced (open one for a preview) and published dashboards and newsletters. View all on either side opens a searchable list, 20 per page.

The Datasets half of the Home resources card listing recently updated datasets
Recently updated datasets. Click a row to open the table.

Analytics

Appears once the workspace has had a run. Tiles for Rows processed, Pipeline runs, Failed runs, Success rate, Average rows / run and Cancelled runs, each with a small trend chart and the change against the previous period. The two filter buttons open Analytics filters, where you choose a Pipeline (default All pipelines) and a Date range (Last 24 hours, Last 7 days, Last 30 days — the default — or Last 90 days).

The Analytics section on Home with metric tiles and trend charts
Workspace analytics for the last 30 days.

Recent chats

Chats are not a Home section, but they are one hover away. Rest the pointer on Chat History in the sidebar and a panel lists New chat and your eight most recent conversations. Chats where Jason is still working also appear under Needs attention.

The sidebar

The left rail is present on every workspace page. It is pinned open by default; the pin button next to the wordmark lets it collapse to icons and expand on hover.

The left sidebar with Home, Chat History, Pipelines, Latest runs, Analytics, Alert Rules, Integrations, the rows-left meter, Search and Settings
The sidebar. Badges appear only for active work.
Item Where it goes Notes
Home / The overview described above.
Chat History Chats Badge counts chats where Jason is working. Hover for recent chats.
Pipelines Pipelines list Under Data collection. Hover for the first ten pipelines, View all and View runs.
Latest runs All runs across pipelines Badge counts running runs.
Analytics Dashboards and newsletters Under Connections. Hover for every dashboard and newsletter.
Alert Rules Alert rules No badge; a count of rules would not be useful.
Integrations Integrations Slack, email, Sheets, webhooks, MCP and the CLI.
rows left Monthly usage dialog See below.
Search Workspace search Also ⌘K (Ctrl+K on Windows and Linux).
Settings Workspace settings Signed-in users only.
Your name Account menu Log out, or Log in when browsing as a guest.

The rows-left meter. The grey card near the bottom reads N rows left with a thin progress bar for the month. It turns amber at 80% of your monthly cap and red when the cap is reached. Click it for the Monthly usage dialog: rows collected, the monthly cap, any once-off row grants, the reset date, and Manage usage →. The full picture is on Billing and usage.

Search (⌘K). Type two or more characters to search datasets, dashboards, pipelines, runs, chats, alerts, reports and integrations by name. With three or more characters you can also Search dataset rows for “…” across the active version of every dataset. When nothing matches, Ask Jason about “…” starts a chat with the question. Arrow keys move, Enter opens, Escape closes.

Jason (⌘J). On every page except a chat page, a Jason button floats bottom-right. ⌘J (Ctrl+J) toggles it. The panel is a compact chat scoped to the page you are on; Open full page moves the same conversation to a chat page. Details on Chatting with Jason.

Above the content, every page has a breadcrumb bar with the workspace name in the centre (click it to switch workspace), and on the right the theme toggle, Share workspace and the notifications bell.

What’s next

Connect your data assistant

Build datasets and work with your data in ChatGPT, Claude, Copilot or another assistant.

Connect in ChatGPT

  1. Open Settings → Security and login and enable Developer mode.
  2. Open Plugins and select + to create a connection. Name it Jsonify, add a short description, and paste the URL below.
  3. Use OAuth for authentication, select Create, and sign in to your Jsonify account when prompted.
  4. Start a new chat and select Jsonify from + → More, then describe your dataset.
Server URLhttps://factory.jsonify.com/mcp

If Developer mode is unavailable, your plan or workspace settings may restrict custom connections.

Official ChatGPT setup guide ↗

Then say: “build me a dataset of competitor product prices and availability, refreshed daily”. Full instructions per client on /connect.