diff --git a/helm/sim/Chart.yaml b/helm/sim/Chart.yaml
index f675d1d7f6b..99aaf25b9ba 100644
--- a/helm/sim/Chart.yaml
+++ b/helm/sim/Chart.yaml
@@ -2,8 +2,8 @@ apiVersion: v2
 name: sim
 description: A Helm chart for Sim - the open-source AI workspace where teams build, deploy, and manage AI agents
 type: application
-version: 1.6.3
-appVersion: "v0.7.44"
+version: 1.6.4
+appVersion: "v0.8.18"
 kubeVersion: ">=1.25.0-0"
 home: https://sim.ai
 icon: https://raw.githubusercontent.com/simstudioai/sim/main/apps/sim/public/logo/primary/primary.svg
diff --git a/helm/sim/README.md b/helm/sim/README.md
index 3bd670f37e5..54ded4b28af 100644
--- a/helm/sim/README.md
+++ b/helm/sim/README.md
@@ -515,6 +515,10 @@ kubectl --namespace sim logs deploy/sim-app -c migrations
 
 ---
 
+## Upgrading to 1.6.4
+
+* `appVersion` — the default image tag for every first-party image (`app`, `realtime`, `migrations`, `pii`, `copilot`) when `image.tag` is unset — moves from `v0.7.44` to `v0.8.18`. It had not been bumped since chart 1.2.0, so an unpinned install deployed an application dozens of releases behind the chart it shipped with, and values keys added by newer charts had no effect because the running image did not read them. **An unpinned release rolls every first-party pod on upgrade.** Installs that pin `image.tag` or `image.digest` are unaffected; pinning explicitly remains the recommendation for production.
+
 ## Upgrading to 1.5.0
 
 Two changes alter behavior on an existing release. Neither requires action, but read both.
diff --git a/helm/sim/ci/kind-values.yaml b/helm/sim/ci/kind-values.yaml
index c3b8464cd8a..28d4f2a7676 100644
--- a/helm/sim/ci/kind-values.yaml
+++ b/helm/sim/ci/kind-values.yaml
@@ -1,7 +1,18 @@
 # CI-only overlay for the kind install test: shrink resource requests so the
 # default configuration schedules on a small CI runner. Layered on top of
 # ci/default-values.yaml. Dummy sizing — never use in a deployment.
+
+# Pin every first-party image to the published :latest rather than letting the
+# tag default to Chart.AppVersion. appVersion names the release the chart ships
+# WITH, and version tags are cut by the main-branch merge commit that releases
+# it (detect-version in ci.yml) — so on the PR that raises appVersion, the tag
+# it names does not exist yet and the install would sit in ImagePullBackOff
+# until --wait times out. That circularity is why appVersion went unbumped from
+# chart 1.2.0 to 1.6.3. This test asks whether the chart installs, not which
+# application build it installs, so any published image answers it.
 app:
+  image:
+    tag: "latest"
   resources:
     requests:
       cpu: 200m
@@ -11,6 +22,8 @@ app:
       memory: 2Gi
 
 realtime:
+  image:
+    tag: "latest"
   resources:
     requests:
       cpu: 50m
@@ -20,6 +33,8 @@ realtime:
       memory: 512Mi
 
 migrations:
+  image:
+    tag: "latest"
   resources:
     requests:
       cpu: 100m
diff --git a/helm/sim/values.yaml b/helm/sim/values.yaml
index 7c8c0910621..2bdcbd71e42 100644
--- a/helm/sim/values.yaml
+++ b/helm/sim/values.yaml
@@ -20,7 +20,7 @@ app:
   # Image configuration
   image:
     repository: simstudioai/simstudio
-    # tag defaults to Chart.AppVersion. Override with a release tag (e.g. "v0.7.44") or pin via image.digest.
+    # tag defaults to Chart.AppVersion. Override with a release tag (e.g. "v0.8.18") or pin via image.digest.
     tag: ""
     # Optional image digest pin: "sha256:..." — when set, overrides tag.
     digest: ""