diff --git a/docs/ROADMAP.md b/docs/ROADMAP.md
index 637bc54..803205e 100644
--- a/docs/ROADMAP.md
+++ b/docs/ROADMAP.md
@@ -33,11 +33,12 @@ Spec: `docs/superpowers/specs/2026-07-01-poker-prompts-design.md`
   departures→TABLE, "stack" leaking questions into LOG, thin MENTAL lexicon). 18
   unit tests. Still a heuristic + swappable seam — upgrade to an LLM/MI50
   classifier only if live misses justify it; keep tuning against real transcripts.
-- 🔨 **Phase C — MI50 tool-calling (code done, flip pending).** Enabling it is now
-  a config toggle: `cfg.tool_backends` (env `TOOL_BACKENDS`, default "cloud"). To
-  turn on: MI50 llama.cpp up with `--jinja` + a tool-capable model, then set
-  `TOOL_BACKENDS="cloud,mi50"`. NOT flipped — MI50 is currently offline, `--jinja`
-  unconfirmed. No code change needed when ready.
+- ✅ **Phase C — MI50 tool-calling (LIVE 2026-07-06).** Added `--jinja` to the
+  lyra-brain llama.cpp launch (`/opt/models/docker-compose.yml` in CT202, so it's
+  reboot-resilient); Qwen2.5-32B confirmed emitting real `tool_calls`. Flipped
+  `TOOL_BACKENDS=cloud,mi50` in `.env`. MI50 chat turns now get the same tool
+  contract as cloud. (Untested in a real poker session on the mi50 backend — worth
+  a live check that tool-calling holds up under the full poker prompt.)
 
 ## Persona (the "person" layer)