DataSocial is an MCP server published by io.github.hashfunction-dev, hosted at api.datasocial.ai. When last checked on 2026-11-21 it was working: it answers and lists its tools. It declares 4 tools, including list_tables, describe_table, examples, run_sql.
Query a TikTok data warehouse (creators, videos, sounds, daily history) with read-only SQL.
1 connections through 5 shared values. Tap any node.
Squares are this server's values; dots are other servers sharing them. A shared address or publisher is a lead, not proof of a common operator.
Inferred from the host name and headers: what a server runs on, not who runs it.
The tables in the DataSocial TikTok warehouse: name, row count, what one row is, and what it holds.
No parameters.
{
"name": "list_tables",
"title": "List tables",
"description": "The tables in the DataSocial TikTok warehouse: name, row count, what one row is, and what it holds.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"properties": {}
},
"outputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"tables"
],
"properties": {
"tables": {
"type": "array",
"items": {
"type": "object",
"required": [
"table",
"rows",
"one_row_is",
"about"
],
"properties": {
"rows": {
"type": "string"
},
"about": {
"type": [
"string",
"null"
]
},
"table": {
"type": "string"
},
"one_row_is": {
"type": [
"string",
"null"
]
}
},
"additionalProperties": false
}
}
},
"additionalProperties": false
},
"annotations": {
"readOnlyHint": true
}
}One table's columns with types and meanings, its sort key and coverage. Read it before writing SQL on a table.
| Parameter | Type | Description |
|---|---|---|
| table * | string | The table, e.g. "creators" or "tiktok.creators" |
{
"name": "describe_table",
"title": "Describe a table",
"description": "One table's columns with types and meanings, its sort key and coverage. Read it before writing SQL on a table.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"table"
],
"properties": {
"table": {
"type": "string",
"description": "The table, e.g. \"creators\" or \"tiktok.creators\""
}
}
},
"outputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"table",
"description"
],
"properties": {
"table": {
"type": "string"
},
"description": {
"type": "string",
"description": "Size, grain, sort key, coverage and every column with its type and meaning"
}
},
"additionalProperties": false
},
"annotations": {
"readOnlyHint": true
}
}Worked examples (a question and SQL that runs here) and the SQL tips: growth maths, history-table patterns, big sound_history days. Pass a table to get only its examples.
| Parameter | Type | Description |
|---|---|---|
| table | string | Only examples on this table, e.g. "sound_history"; omit for all |
{
"name": "examples",
"title": "Examples and SQL tips",
"description": "Worked examples (a question and SQL that runs here) and the SQL tips: growth maths, history-table patterns, big sound_history days. Pass a table to get only its examples.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"properties": {
"table": {
"type": "string",
"description": "Only examples on this table, e.g. \"sound_history\"; omit for all"
}
}
},
"outputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"examples",
"sql_tips"
],
"properties": {
"examples": {
"type": "array",
"items": {
"type": "object",
"required": [
"table",
"question",
"sql"
],
"properties": {
"sql": {
"type": "string"
},
"table": {
"type": "string"
},
"question": {
"type": "string"
}
},
"additionalProperties": false
}
},
"sql_tips": {
"type": "string"
}
},
"additionalProperties": false
},
"annotations": {
"readOnlyHint": true
}
}Run one read-only ClickHouse SELECT on tiktok.* and get the rows back (at most 100 per request; 10 s per query). Name the columns you need and add a LIMIT.
| Parameter | Type | Description |
|---|---|---|
| sql * | string | One ClickHouse SELECT, e.g. SELECT username, followers FROM tiktok.creators WHERE country = 'US' ORDER BY followers DESC LIMIT 10 |
{
"name": "run_sql",
"title": "Run SQL",
"description": "Run one read-only ClickHouse SELECT on tiktok.* and get the rows back (at most 100 per request; 10 s per query). Name the columns you need and add a LIMIT.",
"inputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"sql"
],
"properties": {
"sql": {
"type": "string",
"maxLength": 100000,
"description": "One ClickHouse SELECT, e.g. SELECT username, followers FROM tiktok.creators WHERE country = 'US' ORDER BY followers DESC LIMIT 10"
}
}
},
"outputSchema": {
"type": "object",
"$schema": "http://json-schema.org/draft-07/schema#",
"required": [
"columns",
"rows",
"capped",
"rows_read",
"elapsed_ms"
],
"properties": {
"rows": {
"type": "array",
"items": {
"type": "array",
"items": {}
},
"description": "One array per row, values in the order of columns"
},
"capped": {
"type": "boolean",
"description": "true when the result was cut at the 100-row limit"
},
"columns": {
"type": "array",
"items": {
"type": "string"
}
},
"rows_read": {
"type": "number"
},
"elapsed_ms": {
"type": "number"
}
},
"additionalProperties": false
},
"annotations": {
"readOnlyHint": true,
"openWorldHint": false
}
}DataSocial is a TikTok data warehouse (ClickHouse): billions of creators, videos and sounds, with daily history. Start with list_tables, then describe_table for the columns you need and examples for worked SQL on that table, then run_sql. Rules: one read-only SELECT on tiktok.<table>; name the columns; always a LIMIT; at most 100 rows per request; 10 s per query. Filter the huge tables on their sort key (creators: country, then followers). History tables hold one row per id per date: the latest day is (SELECT max(date) FROM tiktok.<table>). Growth % = (now - previous) / previous * 100. ids are strings (they exceed 2^53). Videos carry no creator_id: a video cannot be linked to its creator. Emails: creators.email is '[email hidden]' when we have one (has_email, email_domain say which); addresses in bios and captions are hidden too. For the 31M creator emails (sold privately) or the full database without the 100-row limit: https://datasocial.ai/buy
Names, descriptions and schemas were written by the server and are shown as data.
Pattern matches on what the server declares. They show what it says it can do, not what it does.
None yet. Changes appear after a server's second daily check.