diff --git a/CHANGELOG.md b/CHANGELOG.md
index 97b2342..e98ceba 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -3,6 +3,9 @@
 ## Unreleased
 
 ### Upgrade Notes
+- **Breaking:** `clawhdf5-agent`'s `agent` feature is removed. It enabled
+  nothing — the agent layer is always built — but the README and guides told
+  people to pass it; drop `agent` from `features = [...]`.
 - **`clawhdf5-migrate` now writes a real agent store.** Its output used to be
   a layout of its own (`/chunks`, `/sessions`, `/entities`, `/relations`, no
   `/meta`) that `HDF5Memory::open` rejected, so a migrated file could not be
diff --git a/README.md b/README.md
index cbf37e9..a800797 100644
--- a/README.md
+++ b/README.md
@@ -615,7 +615,6 @@ ClawhDF5's agent memory design draws from 15+ recent papers:
 | `openblas` | no | OpenBLAS (Linux) |
 | `gpu` | no | GPU search via wgpu |
 | `async` | no | Tokio async with background flush |
-| `agent` | no | Reserved; currently enables nothing (the agent layer is always built) |
 
 To opt out of the parallel build: `--no-default-features --features float16,hnsw`.
 For an exact linear cosine scan instead of HNSW: `--no-default-features --features float16`.
diff --git a/crates/clawhdf5-agent/Cargo.toml b/crates/clawhdf5-agent/Cargo.toml
index 8af2c6f..516e02d 100644
--- a/crates/clawhdf5-agent/Cargo.toml
+++ b/crates/clawhdf5-agent/Cargo.toml
@@ -64,7 +64,6 @@ zstd = ["clawhdf5/zstd"]
 # `--no-default-features` (plus re-enabling other defaults) to force the exact
 # linear cosine scan.
 hnsw = ["clawhdf5-ann"]
-agent = []
 gpu = ["clawhdf5-gpu/gpu-wgpu"]
 fast-math = ["matrixmultiply"]
 accelerate = ["accelerate-src", "cblas-sys"]
diff --git a/docs/QUICKSTART.md b/docs/QUICKSTART.md
index 0635ce0..e8acf84 100644
--- a/docs/QUICKSTART.md
+++ b/docs/QUICKSTART.md
@@ -27,7 +27,7 @@ The core use case. Give your AI agent persistent, searchable memory in a single
 ```toml
 # Cargo.toml
 [dependencies]
-clawhdf5-agent = { version = "2.0", features = ["agent"] }
+clawhdf5-agent = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5" }  # not on crates.io yet
 ```
 
 ### Create a Memory Store
diff --git a/docs/USE_CASES.md b/docs/USE_CASES.md
index 4d0cb4a..132c6f1 100644
--- a/docs/USE_CASES.md
+++ b/docs/USE_CASES.md
@@ -212,7 +212,7 @@ This is the container image for intelligence.
 | Your Situation | Features to Enable | Why |
 |----------------|-------------------|-----|
 | **Quick prototype** | Default | Vector search works out of the box |
-| **Production agent** | `agent`, `float16`, `parallel` | Half-precision saves 50% storage, parallel search for scale |
+| **Production agent** | defaults (`float16`, `hnsw`, `parallel`) | HNSW search and a parallel index build; half-precision *storage* is `MemoryConfig::float16`, on by default for new stores |
 | **macOS** | + `accelerate` | Apple AMX coprocessor for matrix ops |
 | **Linux server** | + `openblas` or `fast-math` | BLAS acceleration |
 | **GPU available** | + `gpu` | wgpu-based search, wins at 100K+ scale |
@@ -220,14 +220,15 @@ This is the container image for intelligence.
 | **Edge device** | Default only | Minimal dependencies, smallest binary |
 
 ```toml
+# Not on crates.io yet: depend on the repository.
 # Production agent on Linux
-clawhdf5-agent = { version = "2.0", features = ["agent", "float16", "parallel", "fast-math"] }
+clawhdf5-agent = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5", features = ["fast-math"] }
 
 # Edge device
-clawhdf5-agent = { version = "2.0", features = ["agent"] }
+clawhdf5-agent = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5" }
 
 # macOS with GPU
-clawhdf5-agent = { version = "2.0", features = ["agent", "float16", "accelerate", "gpu", "async"] }
+clawhdf5-agent = { git = "https://git.redclaw.dev/quantumclaw/clawhdf5", features = ["accelerate", "gpu", "async"] }
 ```
 
 ---