From 2b7541a8b120559e3461b621e442b30cdf009291 Mon Sep 17 00:00:00 2001
From: tobjend <tobend85@gmail.com>
Date: Wed, 1 Jul 2026 11:40:09 +0200
Subject: [PATCH] gif 50% bigger than 37.5% = 56.25%

---
 README.md | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/README.md b/README.md
index bb790f7..339be03 100644
--- a/README.md
+++ b/README.md
@@ -15,7 +15,7 @@ Dervish replaces both with a **one-call MCP tool**: pass your sequences, get bac
 
 **Without Dervish:** token cost scales linearly with examples. **With Dervish:** one compact grammar describes them all — a ~60–200 token rule instead of thousands of tokens of raw examples. Try it out and you too will say:
 
-<p align="center"><img src="dervish.gif" alt="Dervish animation" width="75%"></p>
+<p align="center"><img src="dervish.gif" alt="Dervish animation" width="56.25%"></p>
 
 ## MCP Server
 
@@ -142,7 +142,7 @@ The sweet spot: **multiple implementations of the same abstract task** with a sh
 ## Token savings
 
 <p align="center">
-  <img src="chart_token_savings.png" alt="Token savings per dataset" width="75%">
+  <img src="chart_token_savings.png" alt="Token savings per dataset" width="56.25%">
 </p>
 
 Across all public benchmarks, Dervish delivers **40–83× compression**. The grammar is smaller than a single example file would be — and it represents the entire dataset.