---
title: "How we measure · Docs · SocialToAI"
canonical: https://socialtoai.app/docs/measurements/
updated: 2026-11-20
---

# How we measure

SocialToAI states a figure about itself only with a record of how it was measured. This page prints those records: the date, the sample, the method and the version of the system that was measured.

- Page: https://socialtoai.app/docs/measurements/
- Last updated: 2026-11-20

## 89% fewer tokens across 215 interfaces

| What | Record |
| --- | --- |
| Measured on | 2026-11-06 |
| Sample | Every offered interface with a recorded real answer: one answer each, the platform's raw JSON kept as it came. |
| Method | The same code that serves calls turned each recorded answer into SocialToAI's answer, offline: the JSON an HTTP call and an MCP tool both send, billing and notes included. Both were counted with the o200k tokenizer. No call was made. |
| System version | SocialToAI as served on 2026-11-06 |
| All of them | 8,357,994 tokens raw, 878,622 from SocialToAI |

| Platform | Interfaces | Raw JSON (tokens) | SocialToAI (tokens) | Fewer |
| --- | --- | --- | --- | --- |
| TikTok | 18 | 2,286,105 | 56,180 | 97% |
| Instagram | 8 | 544,022 | 38,520 | 92% |
| X | 21 | 247,096 | 88,867 | 64% |
| Reddit | 11 | 169,102 | 27,114 | 83% |
| YouTube | 10 | 52,538 | 31,967 | 39% |
| Facebook | 4 | 16,144 | 4,526 | 71% |
| LinkedIn | 7 | 36,700 | 21,249 | 42% |
| Threads | 4 | 349,990 | 19,992 | 94% |
| Telegram | 3 | 18,151 | 5,456 | 69% |
| Pinterest | 2 | 43,466 | 4,677 | 89% |
| Xiaohongshu | 21 | 479,829 | 80,827 | 83% |
| Douyin | 33 | 1,836,257 | 99,034 | 94% |
| Weibo | 10 | 543,363 | 25,361 | 95% |
| WeChat Articles | 12 | 51,639 | 19,824 | 61% |
| Bilibili | 8 | 274,162 | 26,909 | 90% |
| Zhihu | 16 | 711,609 | 189,099 | 73% |
| Kuaishou | 8 | 271,296 | 18,647 | 93% |
| WeChat Channels | 11 | 182,863 | 31,561 | 82% |
| Meta Ad Library | 5 | 236,356 | 79,078 | 66% |
| Google Ads Transparency | 3 | 7,306 | 9,734 | -34% |

Where an interface gives only a few numbers, the notes every answer carries (what it cost, the balance, the fields it cannot give) can make it larger than the raw answer; the table counts those too.

## 129×

| What | Record |
| --- | --- |
| Measured on | 2026-10-07 |
| Sample | One keyword search on a short-video platform, the same query sent to a raw platform-data API and to SocialToAI. |
| Method | Both responses were saved as returned and counted in model tokens: about 465,000 for the raw response, about 3,600 for SocialToAI's. 465,000 ÷ 3,600 ≈ 129. |
| System version | SocialToAI as served on 2026-10-07 |
| Raw platform API, one Douyin search | 465,000 tokens |
| SocialToAI | 3,600 tokens |

## 8.0

| What | Record |
| --- | --- |
| Measured on | 2026-10-07 |
| Sample | 7 research questions about public social media, each answered by the same model connected to one product at a time. |
| Method | Same model, same prompt and the same budget for every product, web search off. Each answer was scored from 1 to 10 by a judge model that saw no product names; the figure is the mean over the 7 questions. The other four products are not named. |
| System version | SocialToAI as served on 2026-10-07 |
| SocialToAI | 8.0 |
| Social data MCP A | 7.4 |
| Social data MCP B | 6.1 |
| Social data MCP C | 6.1 |
| Social data MCP D | 4.4 |
