diff --git a/retrieval/README.md b/retrieval/README.md
index 28ca65e..18b8c20 100644
--- a/retrieval/README.md
+++ b/retrieval/README.md
@@ -13,7 +13,7 @@ because keyword-only search misses synonyms.
 | File | Role |
 |------|------|
 | `data.py` | Lazy loader for the arxiv-complete metadata parquet (polars `scan_parquet`). Streams rows so only the matching subset is materialized. |
-| `embed.py` | Pluggable embedding providers: an OpenAI-compatible **remote** provider (production) and a dependency-free **local hashing** fallback (offline, no API key). |
+| `embed.py` | Pluggable embedding providers: an OpenAI-compatible **remote** provider (production), a **local semantic** sentence-transformers model (default local path), and a dependency-free **local hashing** fallback (offline, no deps). |
 | `index.py` | Builds and persists a FAISS flat L2 index + metadata sidecar (paper_id → vector, title, year, category, URL, abstract). Incremental adds. |
 | `query.py` | The interface. Embeds the query, runs ANN search, ranks, applies filters. Falls back to keyword search if no index is available. |
 | `search.py` | CLI tying it together. |
@@ -72,11 +72,11 @@ python search.py --query "speculative decoding" --keyword-only
 python search.py --query "speculative decoding" --survey-only --category cs.LG
 ```
 
-## Embeddings: remote (production) vs local fallback
+## Embeddings: remote (production) vs local semantic vs hashing
 
-This machine (Strix Halo, ~104 GiB unified RAM) has **OOM-killed local models
-before**, so we deliberately do *not* import torch or run a large local
-embedding model here.
+This machine (Strix Halo, ~104 GiB unified RAM) has **OOM-killed large local
+models before**, so we deliberately avoid loading big ones. Three options, all
+behind the same `Embedder` interface:
 
 **Production path** — set these env vars and the CLI uses a remote
 OpenAI-compatible embeddings endpoint:
@@ -90,30 +90,21 @@ export EMBEDDING_MODEL=text-embedding-3-small
 Any endpoint with the OpenAI embeddings shape works (OpenAI, Together, Groq, a
 local vLLM/Ollama server, …).
 
-**Offline fallback** — with no key configured, the code automatically uses
-`HashingEmbedder`, a deterministic, dependency-free embedding that projects a
-bag of character n-grams into a fixed-size vector via the hashing trick. It needs
-no weights, no GPU, no internet and no API key, so the whole ANN pipeline is
-fully exercisable for demos and CI. (It is not semantically meaningful — synonyms
-are not mapped together — but it shares enough lexical overlap to cluster near
-the right topics.)
+**Local semantic (default)** — with no API key configured, the code
+automatically uses `LocalEmbedder`, a small CPU
+[sentence-transformers](https://huggingface.co/sentence-transformers) model
+(`paraphrase-MiniLM-L3-v2`, ~90 MB). This produces *genuinely semantic* vectors
+(synonyms and paraphrases map close together) at no API cost, and fits comfortably
+in RAM. It's slower than keyword hashing — embedding the full 3.15M corpus takes
+~1-2 hours on this box, so for a demo index a single `--category` subset instead.
+Override the model with `EMBEDDING_LOCAL_MODEL` (e.g. `all-MiniLM-L6-v2` for
+higher quality at ~3x the time).
 
-### Swapping in a real local model (if you have the RAM)
-
-Add a class in `embed.py`:
-
-```python
-class LocalEmbedder(Embedder):
-    DIM = 768
-    def __init__(self):
-        from sentence_transformers import SentenceTransformer
-        self._model = SentenceTransformer("all-MiniLM-L6-v2")
-    def embed(self, texts):
-        return self._model.encode(texts, normalize_embeddings=True)
-```
-
-and pass it via `embedder=LocalEmbedder()` to the index/query builders. The
-rest of the pipeline is agnostic to where the vectors come from.
+**Dependency-free fallback** — pass `--embedder hashing` (or `--keyword-only`)
+for `HashingEmbedder`, a deterministic embedding that projects a bag of character
+n-grams into a fixed-size vector via the hashing trick. No weights, no GPU, no
+internet, no API key — handy for CI or a smoke test, but not semantically
+meaningful (synonyms aren't mapped together).
 
 ## How the index works
 
diff --git a/retrieval/requirements.txt b/retrieval/requirements.txt
index 6b39865..d2d5609 100644
--- a/retrieval/requirements.txt
+++ b/retrieval/requirements.txt
@@ -13,6 +13,10 @@ numpy
 polars
 faiss-cpu
 
+# Needed for the local semantic embedder (LocalEmbedder, the default local path).
+# Uses a small CPU sentence-transformers model (~90 MB, no GPU needed).
+sentence-transformers
+
 # Only needed for the remote (production) embedding provider.
-# Comment out if you only use the local HashingEmbedder fallback.
+# Comment out if you only use the local embedders.
 requests