Why Era

One seeded graph, projected into every vendor's schema.

The account behind a Salesforce opportunity is the account a Zendesk ticket and a Gong call reference, because each system is a projection of one company graph rather than a table of its own.
The graph is generated from a seed, a clock and a day-state - 730 days of trading, hiring and filing - and the same three inputs rebuild it down to the ids.

The method

What the generator does that a row filler does not.

rates, not countsNo collection is given a size. The desk is given a rate - fifteen tickets per customer per year - and then capped by the people who have to work them, at nine closes per agent per day. Files come off headcount, paperwork off the book, recorded calls off the reps who could have taken the meetings. Ask for a bigger company and you get a bigger business, not the same estate with a wider id range.
a working calendarEvery timestamp is sampled in the record's own timezone against an hour-of-day curve: two peaks, a lunch dip, a thin evening tail, weekday weighting and almost nothing on a public holiday. Volume ramps with the company's growth, so a group-by-month reads like a company that grew rather than a window filled at random.
lifecycles, not end statesA closed deal is the walk that got it there - stage by stage, each with its own dwell, a median cycle of about eight weeks and a long one for the enterprise account. A solved ticket was responded to before it was resolved, so "what changed between March and June" has something to read.
people with a jobThe average customer knows six named people, spent across the book by size: the champion, the economic buyer, the security reviewer and the admins who file the tickets. Employees have tenure, teams, work history and time off, and the agent who answered a ticket is one of them.
mess, declaredDuplicate accounts, alias emails, pasted formatting and blank optional fields are an opt-in overlay, and every defect is written to a manifest - kind, field, record, before and after - so an eval can score whether the agent found the duplicates instead of guessing.
one inputA seed, a clock and a day-state. Give the generator the same three and you get the same company back, down to the ids.
One graph, every schema

Projection, not a generator per system.

The entities - people, customers, deals, tickets, calls, files, chat - reference each other in the graph. A connector adds no data of its own: it maps those entities onto its vendor's resource shapes, pagination and error codes, which is why the ids still line up once the systems are apart.

One seeded graph customers deals tickets calls messages files people the company
projected from the company records that name each other

The dashed ring is the part a per-system generator cannot have: neighbours holding each other's ids, so a cross-system join resolves.

Each vendor's own schema
SalesforceAccount · Contact · Opportunity · Task
ZendeskOrganization · User · Ticket · Comment
SlackChannel · Message · File
GongCall · Transcript · Interaction stats
HiBobEmployee · Work history · Time off
S3Bucket · Object · Version
and the rest of the fleet

The requester on a Zendesk ticket is a contact at the account the Salesforce opportunity names, and the agent who answered is the employee HiBob has on that team. The ids match because there is one company underneath, not three.

Depth

Two years of it, at the scale you ask for.

History runs 730 days back, because a year-over-year question and an annual renewal both need last year to exist. Scale is a choice at build time - the vocabulary this console builds in - and the rows below are measured from the systems rather than modelled on this page.

TierThe companyRows, four measured systems
Measuring…
consumer scaleA consumer book runs to millions of customers, with the queue, the events and the recorders sized off the people who could have handled them.
across the fleetThe same company is written into every system you pick from this deployment's catalog, one isolated tenant with its own token.
Moving in time

The same company, a quarter later.

Generating again with another seed gives a different company, which is no use to anything that syncs. Era holds three states of the same one, and the quarter is derived once for the graph so every system moves by the same thirteen weeks.

day0Inception. The tenant stands up empty, which is what a first full sync should be pointed at.
day1Today. Two years of trading, hiring, filing, talking and closing, already behind the company.
day291 days on, ending in a quarter close: accounts that signed, people who joined and left, deals that advanced, the tickets, calls and files of thirteen weeks. "Now" moves with it, so a dashboard window and a last-activity date read as a quarter later rather than as today with rows from the future bolted on.
Testing & CI · day-states, faults, tenants
Proof

Scored by a scorer, checked by the vendors' own clients.

Every tenant is scored against reference distributions and swept by an adversarial detector looking for the tells of generated data, and the baseline a CI gate holds the generator to is published from the scorer's own output.

The shape of an API is checked the same way, with code nobody here wrote. zenpy's Zendesk suite runs with its cassettes flipped to record mode, putting requests it once recorded against a real account back on the wire. psycopg 3's suite is written against a server rather than the driver, so it can be pointed at the Postgres simulator. Then stripe-python, the Jira client, elasticsearch-py, minio-py, pymysql, clickhouse-connect and the vitest suites from vercel-labs/emulate. Every upstream is pinned, and every connector carries an audit file saying what was probed, what was broken and which test pins the fix.

Change the base URL and the vendor's SDK talks to it · each system says what it is at /_meta
Where it does not fit

What Era is not.

Every environment is 100% simulated: no real records, no real people, and no substitute for a staging system carrying your customers' data. Each system is vendor-shaped, not the vendor: it is not affiliated with, endorsed by or sponsored by any product it models, and it implements the surfaces a client uses rather than every corner of an API. Read what a simulator is not before you let a released assertion depend on one.