↗ free · no signup · works in 30 seconds

smart-mock.com

Realistic fake data. Live REST API. Yours instantly.

Create a table, pick your columns, get a URL you can call from any app - curl, Python, JavaScript, whatever. Relational data, error injection, custom value lists. Free forever.

or let your AI build it for you ↓

→ get started free try it out, no signup

LIVE EXAMPLE

PRD-0042 · Wireless Headphones · category: electronics · price: 89.99 · stock: 143 → ORD-0019 → PRD-0042 · qty: 2 · total: 179.98 · status: shipped → USR-0007 · plan: pro · created: 2024-11-03 · city: Berlin → SES-0291 → USR-0007 · duration: 847s · device: mobile · events: 12 → PLR-0012 · username: dataset(handles) → xShadow99 · level: 34 · score: 88420 → RUN-0091 → AGT-0003 · model: gpt-4o · tokens: 1240 · status: done → PRD-0042 · Wireless Headphones · category: electronics · price: 89.99 · stock: 143 → ORD-0019 → PRD-0042 · qty: 2 · total: 179.98 · status: shipped → USR-0007 · plan: pro · created: 2024-11-03 · city: Berlin → SES-0291 → USR-0007 · duration: 847s · device: mobile · events: 12 → PLR-0012 · username: dataset(handles) → xShadow99 · level: 34 · score: 88420 → RUN-0091 → AGT-0003 · model: gpt-4o · tokens: 1240 · status: done →

// built-in types + your own

What gets generated.

first_nameJohnrule:first_name

middle_nameMichaelrule:middle_name

last_nameCarterrule:last_name

email[email protected]rule:email

address742 Market St, San Franciscorule:address

date2025-03-14rule:date

number3,206.39range(1000–8000)

booleantruerule:bool

uuida3f7-…-c12brule:uuid

enumbronze / silver / goldrule:enum(…)

datasetAlice · Daniel · Sofiayour list

composedACC-001 … ACC-999generated IDs

// built-in datasets

first_name middle_name last_name email address city country company job_title

// your custom datasets

dataset(agents) dataset(regions) dataset(statuses) + up to 100 new per user

smart-mock · live table universe

suppliers 12 rows

idSUP-0007

nameTechCorp

countryGermany

products 8 rows

supplierTechCorp

idPRD-0042

nameHeadphones

price89.99

orders 30 rows

productTechCorp - Headphones

qty2

total179.98

// suppliers → products → orders - three levels, two cables

// rule nodes - computed columns at generation time

                IF node
                
              
if price > 100
→ "required"
elif price > 20
→ "optional"
else
→ "none"

                products8 rows
              
                name
                Headphones
              
                price
                89.99
              
                insurance
                optional

// what it does

Everything you need to fake it.

⌗

Schema DSL

Define columns with a simple text format. Types, rules, ranges - all in one line.

→ column definition

"email" : "string, rule:email"
"salary": "number, rule:decimal,
           range(1000-8000)"
"joined": "date, rule:from(2020-01-01)
           :until(2025-12-31)"
"tier"  : "string,
           rule:enum(bronze|silver|gold)"

⇌

Bind & Drift

Same entity, same email. Always. Or add drift for controlled inconsistency.

→ consistent entities

// customers table
"email": "string, rule:email,
          bind:cust_email"

// messages table - same email
"email": "string, rule:email,
          bind:cust_email"

// payments - 5% chance different
"email": "string, rule:email,
          bind:cust_email, drift(0.05)"

⊗

Error Injection

Simulate null values, wrong types, malformed dates. Set error_rate 0.0 – 1.0.

→ messy data on demand

// clean data
{ "error_rate": 0.0 }

// 10% bad rows
{ "error_rate": 0.1 }

// chaos mode
{ "error_rate": 1.0 }

// what gets injected:
// null · wrong_type · bad_format
// out_of_range · swapped columns

⊛

Datasets

Create reusable value lists. Reference them in any table. Max 15 per user.

→ custom value lists

POST /users/marek/datasets
{
  "name": "agents",
  "values": ["Alice","Bob","Fero"]
}

// use in column:
"agent": "string,
          rule:dataset(agents)"

// composed IDs:
{ "values": ["ACC-", 1, 999],
  "composed": true }
→ ACC-1, ACC-2 ... ACC-999

⊕

Bounded Generate

Link child records to parent keys. 3 invoices per customer, automatically.

→ relational mock data

POST /users/marek/tables
{
  "table": "invoices",
  "bounded": {
    "ref_table": "customers",
    "ref_key": "",
    "rows": 3
  }
}
// → 3 invoices × each customer
// customers_key auto-populated

∿

Rule Nodes

Computed columns at generation time. Math, sequences, IF conditions - wired into table columns.

→ computed columns

// running balance per account
type: sequence
source: 50, step: 10
group_by: account_id

// conditional status
type: if
if amount > 1000
→ status = "high_value"
else → status = "normal"

◎

Permanent Snapshot

One stable URL per account. SQL query across all tables. Call it like a read-only database.

→ stable API + SQL

// permanent URL
GET /users/{id}/snapshot
    /tables/orders?token=...

// SQL across all tables
POST /snapshot/query
{
  "sql": "SELECT * FROM orders
    JOIN customers
    ON orders.customer_key
       = customers.key"
}

⟡

AI Agent Ready

Give an AI agent your smai_ token. It builds your full schema - tables, cables, rule nodes - via API.

→ schema via API

// user generates smai_ token
// AI agent gets token + user_id
// AI calls:

POST /tables       → customers
POST /tables       → orders
POST /table-links  → wire FK
POST /rule-nodes   → compute

// 5 min TTL · create-only
// no read · no delete

// built for AI agents

Let your AI build it.

Generate an smai_ token and give it to your AI. It can call the API directly - or generate curl commands you run yourself. Either way, your schema is ready in seconds.

Works with any AI that can write HTTP calls - ChatGPT, Claude, Cursor, n8n, LangChain. The token is create-only for row data - it can read schema structure to build on top of it, but cannot read, modify, or delete your generated data.

→ get your smai_ token

AI agent - building schema

// option A: paste token into your AI chat
// → AI generates curl commands, you run them

// option B: give token to an agent with tools
// → agent calls the API directly

// either way, agent calls:

POST /users/{user_id}/tables
Authorization: Bearer smai_xxxxxxxxxxxx
→ "customers" table created

POST /users/{user_id}/table-links
→ orders.customer_id wired to customers.id

// schema ready. start fetching.

// available from token tier (username only, no email) · 5 min TTL · 20 calls · no row data access · no deletes

// no account needed

Try it out.

Jump straight into the workspace - no signup, no username. Guest sessions are temporary. Sign up later to keep your data.

⚠ guest mode - data is not persisted between sessions

          As a guest you can:

          ✓ create tables and generate mock data

          ✓ use the API with a guest token

          ✓ export data via curl

          ~ session lasts 2 hours

          ✗ no snapshots, no saved datasets

smart-mock api · live

// no setup needed - click open workspace →

// then create a table:
POST /users/{guest_id}/tables

// get live mock data on every call:
GET /users/{guest_id}/tables/{table}/live

// example response:
[
{ "id": "CUS-0001", "name": "Alice" },
{ "id": "CUS-0002", "name": "Bob" }
]

_

smart-mock.com

What gets generated.

Everything you need to fake it.

Let your AI build it.

Try it out.

How it works.