diff --git a/internal/lua/adapters/opencode.lua b/internal/lua/adapters/opencode.lua
index 9690b87..db588d9 100644
--- a/internal/lua/adapters/opencode.lua
+++ b/internal/lua/adapters/opencode.lua
@@ -13,15 +13,39 @@ adapter.headers = {
 }
 
 -- OpenAI /chat/completions format (pass-through, strip provider-specific fields)
+-- zen 上游 schema 只接受 text content part（无视觉/音频能力）：多模态 part
+-- （image_url / input_audio / file 等）一律剥离；因此失去全部 content 的
+-- 消息整条丢弃，避免上游 "unknown variant `image_url`, expected `text`"。
 function adapter.transform_request(raw_body)
     local ok, req = pcall(json.decode, raw_body)
     if not ok then return raw_body end
     req.disable_thinking = nil
     req.extra_body = nil
     if req.messages then
+        local kept = {}
         for _, msg in ipairs(req.messages) do
             msg.reasoning_content = nil
+            local drop = false
+            if type(msg.content) == "table" then
+                local parts = {}
+                for _, part in ipairs(msg.content) do
+                    if type(part) == "table" and part.type ~= nil and part.type ~= "text" then
+                        -- multimodal part not supported by zen
+                    else
+                        table.insert(parts, part)
+                    end
+                end
+                if #parts == 0 then
+                    drop = true
+                else
+                    msg.content = parts
+                end
+            end
+            if not drop then
+                table.insert(kept, msg)
+            end
         end
+        req.messages = kept
     end
     return json.encode(req)
 end
diff --git a/internal/lua/vm_test.go b/internal/lua/vm_test.go
index 3e4c3d4..67e8f6c 100644
--- a/internal/lua/vm_test.go
+++ b/internal/lua/vm_test.go
@@ -135,6 +135,37 @@ func TestOpenCodeAdapterFingerprint(t *testing.T) {
 	}
 }
 
+func TestOpenCodeStripsMultiModalParts(t *testing.T) {
+	vm := NewVM(freshAdapterDir(t))
+	if err := vm.Start(); err != nil {
+		t.Fatal(err)
+	}
+	defer vm.Stop()
+	body := `{"model":"x","messages":[
+		{"role":"user","content":[
+			{"type":"text","text":"describe"},
+			{"type":"image_url","image_url":{"url":"data:image/png;base64,QUJD"}},
+			{"type":"input_audio","input_audio":{"data":"QQ","format":"wav"}}
+		]},
+		{"role":"user","content":[{"type":"image_url","image_url":{"url":"https://ex.com/a.png"}}]},
+		{"role":"assistant","content":"plain"}
+	]}`
+	out, err := vm.Transform("opencode", "transform_request", body)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if strings.Contains(out, "image_url") || strings.Contains(out, "input_audio") {
+		t.Fatalf("opencode must strip multimodal parts: %s", out)
+	}
+	if !strings.Contains(out, "describe") {
+		t.Fatalf("text parts must survive: %s", out)
+	}
+	// image-only message dropped: only the text message and plain assistant message remain
+	if got := strings.Count(out, `"role"`); got != 2 {
+		t.Fatalf("image-only message must be dropped, got %d messages: %s", got, out)
+	}
+}
+
 func TestBuildHeadersCustomHook(t *testing.T) {
 	vm := NewVM(freshAdapterDir(t))
 	if err := vm.Start(); err != nil {