diff --git a/generate/__init__.py b/generate/__init__.py
index 03207c4b..6a9a5db2 100644
--- a/generate/__init__.py
+++ b/generate/__init__.py
@@ -13,7 +13,7 @@ from .dialogue import (
     propose_dialogue,
     trajectory_blade,
 )
-from .stream import generate, agenerate
+from .stream import generate
 from .surface import SentenceAssembler, SentencePlan, assemble as assemble_surface
 
 __all__ = [
@@ -23,7 +23,6 @@ __all__ = [
     "FrameSlot",
     "Proposition",
     "PropositionFrame",
-    "agenerate",
     "blade_alignment",
     "classify_dialogue_blade",
     "generate",
diff --git a/generate/stream.py b/generate/stream.py
index 56b3360b..a4f51a70 100644
--- a/generate/stream.py
+++ b/generate/stream.py
@@ -638,48 +638,3 @@ def generate(
         admissibility_trace=tuple(admissibility_trace),
         region_was_unconstrained=region_was_unconstrained,
     )
-
-
-async def agenerate(
-    state: FieldState,
-    vocab,
-    persona,
-    max_tokens: int = 128,
-    vault=None,
-    recall_top_k: int = 3,
-):
-    current = state
-    recent_nodes = deque([state.node], maxlen=_RECENT_WINDOW)
-    stop_nodes = frozenset(
-        idx for token in _STOP_TOKENS
-        if (idx := _try_index(vocab, token)) is not None
-    )
-    for _ in range(max_tokens):
-        current, _hits_applied = _recall_state(
-            _voiced_state(current, persona),
-            vault,
-            recall_top_k,
-        )
-        word, word_idx = _nearest_next(
-            vocab,
-            current.F,
-            current.node,
-            recent_nodes=tuple(recent_nodes),
-            stop_nodes=stop_nodes,
-        )
-        yield _articulate(vocab, word)
-
-        A = vocab.get_versor_at(current.node)
-        B = vocab.get_versor_at(word_idx)
-        V = word_transition_rotor(A, B)
-
-        current = propagate_step(current, V)
-        current = FieldState(
-            F=current.F,
-            node=word_idx,
-            step=current.step,
-            holonomy=current.holonomy,
-            energy=current.energy,
-            valence=current.valence,
-        )
-        recent_nodes.append(word_idx)