Initial commit: standalone vector memory server

REST API for per-agent semantic memory retrieval.

- FastAPI server with /api/{agent}/save and /api/{agent}/query
- ChromaDB for persistent vector storage
- all-MiniLM-L6-v2 via sentence-transformers for embeddings
- Per-agent collections for clean separation
- Config through env vars or config.py
- .venv ready with all dependencies

config.py

@@ -0,0 +1,17 @@
+"""Configuration for the Vector Memory Server."""
+
+import os
+
+
+class Config:
+    DATA_DIR = os.environ.get(
+        "MEMORY_SERVER_DATA",
+        "/home/admin/agent-dir/vector_memory"
+    )
+    EMBEDDING_MODEL = os.environ.get(
+        "MEMORY_SERVER_MODEL",
+        "all-MiniLM-L6-v2"
+    )
+    DEFAULT_TOP_N = int(os.environ.get("MEMORY_SERVER_TOP_N", "5"))
+    HOST = os.environ.get("MEMORY_SERVER_HOST", "127.0.0.1")
+    PORT = int(os.environ.get("MEMORY_SERVER_PORT", "8000"))