# Twigg — AI Compare Ninja

> Hosted context store, assembler, and model router: create a chat once, send only the next event.

## Overview
- **Category:** Development & Code (Agent & RAG frameworks)
- **Status:** Active · checked - Active.
- **Paid from:** 5% / per 1M tokens (Prepaid usage)
- **Free access:** No free plan published
- **Website:** https://twigg.ai/
- **Canonical URL:** https://www.aicompare.ninja/en/tool/twigg

## What it does
- Context stays with the host, model choice per turn
- Token rates include a 5% margin, prepaid instead of a seat SKU

## In brief
Twigg stores the chat, fits context to each model window, and routes to Anthropic, OpenAI, Google, xAI, Fireworks, or OpenRouter. The app sends only the next event, not the transcript. This is not OpenRouter alone.

Twigg bills USD per 1M tokens including a 5% platform margin. Samples: claude-fable-5-1 $10.5 / $52.5, gpt-6-astra $10.5 / $52.5 (above 272k: $21 / $78.75), claude-sonnet-5 $2.1 / $10.5, gpt-5-6-luna $0.21 / $1.26. Empty balance returns 402.


## Limits
HTTP 402 payment_required when the balance cannot cover the turn's input. No SDK; SSE stream. Twigg never executes a tool. Many more Fireworks and OpenRouter rows are available through Twigg.

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Prepaid usage | 5% | per 1M tokens | 5% platform margin over provider cost, prepaid balance; HTTP 402 when the balance cannot cover the turn; no seat SKU |
| claude-fable-5-1 | $10.5 / $52.5 | per 1M tokens | Input $10.5, output $52.5 per 1M tokens including 5% margin; Anthropic, 1M context; cache read $0.2625, cache write $13.125 |
| gpt-6-astra | $10.5 / $52.5 | per 1M tokens | Input $10.5, output $52.5 per 1M tokens including 5% margin; Above 272k context: $21 / $78.75 for the whole request |
| claude-sonnet-5 | $2.1 / $10.5 | per 1M tokens | Input $2.1, output $10.5 per 1M tokens including 5% margin; Anthropic, 1M context |
| gpt-5-6-luna | $0.21 / $1.26 | per 1M tokens | Input $0.21, output $1.26 per 1M tokens including 5% margin; OpenAI, 1M context |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://api.twigg.ai/v1/catalogue/models.md
- https://twigg.ai/docs/quickstart.md
- https://twigg.ai/

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/twigg*