diff --git a/memory-systems/agent-os/04_projects/ai-talk-show/status.md b/memory-systems/agent-os/04_projects/ai-talk-show/status.md
new file mode 100644
index 0000000..b7e413c
--- /dev/null
+++ b/memory-systems/agent-os/04_projects/ai-talk-show/status.md
@@ -0,0 +1,34 @@
+# AI Talk Show — status
+
+**Updated:** 2026-07-09  
+**Repo:** `Tony_tech/ai-talk-show`  
+**Local:** `/home/vpsadmin/projects/ai-talk-show/`  
+
+## What it is
+
+Two practitioner agents (Leonard + Charlie), Moltbook-style dialogue on AI/agent topics, rendered to audio. Target evolution: 7–9 min eps, human-utility lens, producer ratings feedback.
+
+## Done
+
+- Episodes 001–008 transcripts/outputs pattern  
+- System prompts + CORE RULES (self-knowledge carve-out) on Gitea  
+- Producer briefs 004–007  
+- TTS pipeline: **edge-tts** primary (Guy/Eric)  
+- Chatterbox weights/integration explored on VPS then **cancelled for VPS**  
+
+## Parked / off agenda
+
+- **VPS native Chatterbox + swap** — do not pursue  
+- If native TTS wanted: **Tony's laptop**, not this VPS  
+
+## Next (product, not infra)
+
+- Prompt upgrade: domain expertise + benefits/signal-noise for human listeners (when Tony green-lights)  
+- SHOW_NOTES + ratings loop  
+- Length cap ~7–9 min for new pilots  
+- MoA only if Tony requests design debate  
+
+## Runtime notes
+
+- Show via `run_show.py` (not orchestrator as primary)  
+- Model endpoint free-tier oriented; no Grok for agent runs