feat(provider): add Chroma built-in embedding requester

Add chromaembed.py using Chroma's DefaultEmbeddingFunction (all-MiniLM-L6-v2)
for local embedding generation via ONNX Runtime. Also simplify seekdbembed.py
and add ndarray-to-list conversion for JSON serialization compatibility.

src/langbot/pkg/provider/modelmgr/requesters/chroma.svg

@@ -0,0 +1,7 @@
+<svg width="24" height="24" viewBox="0 0 24 24" fill="none" xmlns="http://www.w3.org/2000/svg">
+  <rect width="24" height="24" rx="5" fill="#7B68EE"/>
+  <circle cx="12" cy="12" r="6" fill="#FF6B35"/>
+  <circle cx="12" cy="12" r="3" fill="#7B68EE"/>
+  <path d="M12 6V18" stroke="#FFF" stroke-width="1.5" stroke-linecap="round"/>
+  <path d="M6 12H18" stroke="#FFF" stroke-width="1.5" stroke-linecap="round"/>
+</svg>

src/langbot/pkg/provider/modelmgr/requesters/chromaembed.py

@@ -0,0 +1,61 @@
+from __future__ import annotations
+
+import typing
+
+from .. import requester
+
+REQUESTER_NAME: str = 'chroma-embedding'
+
+
+class ChromaEmbedding(requester.ProviderAPIRequester):
+    """Chroma built-in embedding requester.
+
+    Uses chromadb's DefaultEmbeddingFunction (all-MiniLM-L6-v2).
+    The embedding function runs locally using ONNX Runtime.
+    """
+
+    default_config: dict[str, typing.Any] = {
+        'base_url': '',
+    }
+
+    _embedding_function = None
+
+    async def initialize(self):
+        try:
+            from chromadb.utils import embedding_functions
+        except ImportError:
+            raise ImportError('chromadb is not installed. Install it with: pip install chromadb')
+
+        self._embedding_function = embedding_functions.DefaultEmbeddingFunction()
+
+    async def invoke_llm(
+        self,
+        query,
+        model: requester.RuntimeLLMModel,
+        messages: typing.List,
+        funcs: typing.List = None,
+        extra_args: dict[str, typing.Any] = {},
+        remove_think: bool = False,
+    ):
+        raise NotImplementedError('Chroma embedding does not support LLM inference')
+
+    async def invoke_embedding(
+        self,
+        model: requester.RuntimeEmbeddingModel,
+        input_text: typing.List[str],
+        extra_args: dict[str, typing.Any] = {},
+    ) -> typing.List[typing.List[float]]:
+        """Generate embeddings using Chroma's DefaultEmbeddingFunction."""
+        if self._embedding_function is None:
+            await self.initialize()
+
+        try:
+            result = self._embedding_function(input_text)
+            # DefaultEmbeddingFunction returns list of ndarray, convert for JSON
+            if isinstance(result, list):
+                return [item.tolist() if hasattr(item, 'tolist') else item for item in result]
+            return result.tolist() if hasattr(result, 'tolist') else result
+        except Exception as e:
+            from .. import errors
+
+            raise errors.RequesterError(f'Chroma embedding failed: {str(e)}')

src/langbot/pkg/provider/modelmgr/requesters/chromaembed.yaml

@@ -0,0 +1,21 @@
+apiVersion: v1
+kind: LLMAPIRequester
+metadata:
+  name: chroma-embedding
+  label:
+    en_US: Chroma Embedding
+    zh_Hans: Chroma 嵌入
+  description:
+    en_US: Chroma built-in embedding model (all-MiniLM-L6-v2), runs locally using ONNX Runtime. First-time use will download model files automatically.
+    zh_Hans: 使用 Chroma 内置嵌入模型 (all-MiniLM-L6-v2)，基于 ONNX Runtime 本地运行。首次使用时将自动下载模型文件。
+    ja_JP: Chroma 組み込み埋め込みモデル (all-MiniLM-L6-v2) を使用します。ONNX Runtime でローカル実行。初回使用時にモデルファイルが自動ダウンロードされます。
+  icon: chroma.svg
+spec:
+  config: []
+  support_type:
+  - text-embedding
+  provider_category: builtin
+execution:
+  python:
+    path: ./chromaembed.py
+    attr: ChromaEmbedding

src/langbot/pkg/provider/modelmgr/requesters/seekdbembed.py

@@ -46,14 +46,15 @@ class SeekDBEmbedding(requester.ProviderAPIRequester):
         extra_args: dict[str, typing.Any] = {},
     ) -> typing.List[typing.List[float]]:
         """Generate embeddings using SeekDB's built-in embedding function."""
+        if self._embedding_function is None:
+            await self.initialize()
+
         try:
-            if self._embedding_function is None:
+            result = self._embedding_function(input_text)
-                await self.initialize()
+            # Ensure JSON serialization compatibility
-
+            if isinstance(result, list):
-            if self._embedding_function is None:
+                return [item.tolist() if hasattr(item, 'tolist') else item for item in result]
-                raise RuntimeError('SeekDB embedding function initialization failed')
+            return result.tolist() if hasattr(result, 'tolist') else result
-
-            return self._embedding_function(input_text)
         except Exception as e:
             from .. import errors