Pablo Claude Opus 4.7 (1M context) commited on
Commit
4249091
·
1 Parent(s): 68ba0f8

docs: embed demo video in README · live 5-agent pipeline run

Browse files

Adds assets/video_live.mp4 (5.2 MB) and a <video> tag at the top of the
Live Demo section pointing to the raw GitHub URL — GitHub renders this
inline when the README loads. The static dashboard_live_demo.png stays
as the poster + fallback frame for browsers that block autoplay or for
RSS / mirror viewers that don't render <video>.

The hero TOC now links the demo video first.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Files changed (2) hide show
  1. README.md +21 -1
  2. assets/video_live.mp4 +3 -0
README.md CHANGED
@@ -22,7 +22,7 @@
22
  <p align="center">
23
  <a href="#-the-problem">Problem</a> ·
24
  <a href="#-the-solution">Solution</a> ·
25
- <a href="#-live-demo">Live Demo</a> ·
26
  <a href="#-benchmark-results">Benchmarks</a> ·
27
  <a href="#-architecture">Architecture</a> ·
28
  <a href="#-quick-start">Quick Start</a> ·
@@ -84,6 +84,26 @@ Every optimization traces back to a peer-reviewed paper published at **NeurIPS,
84
 
85
  Real metrics from `demo/app.py` running against the full ContextForge stack — five agents, real Qwen3 tokenizer, real LSH+FAISS dedup, INV-15 enforced live. Side-by-side comparison: **263 → 53 tokens, 79.85 % savings** with ContextForge; passthrough on the right.
86
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
87
  <p align="center">
88
  <img src="assets/screenshots/dashboard_live_demo.png" alt="Live Demo — With vs Without ContextForge, 79.85% savings, INV-15 firing on the Critic" width="960"><br>
89
  <em>Live Demo tab — left: query input. Right: <b>With ContextForge</b> (79.85 % savings, INV-15 fires on the Critic) vs. <b>Without ContextForge</b> (passthrough, 0 % savings).</em>
 
22
  <p align="center">
23
  <a href="#-the-problem">Problem</a> ·
24
  <a href="#-the-solution">Solution</a> ·
25
+ <a href="#%EF%B8%8F-demo-video"><b>▶️ Demo video</b></a> ·
26
  <a href="#-benchmark-results">Benchmarks</a> ·
27
  <a href="#-architecture">Architecture</a> ·
28
  <a href="#-quick-start">Quick Start</a> ·
 
84
 
85
  Real metrics from `demo/app.py` running against the full ContextForge stack — five agents, real Qwen3 tokenizer, real LSH+FAISS dedup, INV-15 enforced live. Side-by-side comparison: **263 → 53 tokens, 79.85 % savings** with ContextForge; passthrough on the right.
86
 
87
+ ### ▶️ Demo video
88
+
89
+ <p align="center">
90
+ <video src="https://github.com/SuarezPM/Apohara_Context_Forge/raw/main/assets/video_live.mp4"
91
+ controls
92
+ muted
93
+ playsinline
94
+ width="900"
95
+ poster="assets/screenshots/dashboard_live_demo.png">
96
+ Your browser does not render embedded video.
97
+ <a href="https://github.com/SuarezPM/Apohara_Context_Forge/raw/main/assets/video_live.mp4">Download the demo video</a>.
98
+ </video>
99
+ </p>
100
+
101
+ <p align="center">
102
+ <em>End-to-end run: query → 5-agent pipeline → 79.85 % token savings → JCR Safety Gate fires INV-15 on the Critic.</em>
103
+ </p>
104
+
105
+ ### Static frame
106
+
107
  <p align="center">
108
  <img src="assets/screenshots/dashboard_live_demo.png" alt="Live Demo — With vs Without ContextForge, 79.85% savings, INV-15 firing on the Critic" width="960"><br>
109
  <em>Live Demo tab — left: query input. Right: <b>With ContextForge</b> (79.85 % savings, INV-15 fires on the Critic) vs. <b>Without ContextForge</b> (passthrough, 0 % savings).</em>
assets/video_live.mp4 ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a7d99c55d7f06af40a6997c3cbf836f3bd192f2407065d5fa1e81fb4be5ee621
3
+ size 5353802