my prefect server setup
prefect-metrics.waow.tech
python
orchestration
use haiku for curation agent — sonnet rate limits on tool-heavy loops
separate curate-model variable (default claude-haiku-4-5) so curation
doesn't share the morning-model config or hit sonnet's 30k/min input
token limit during multi-turn tool calls.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
+1
-1
+1
-1
flows/curate.py
+1
-1
flows/curate.py
···
520
520
openai_key = (await Secret.load("openai-api-key")).get()
521
521
bsky_handle = (await Secret.load("atproto-handle")).get()
522
522
bsky_password = (await Secret.load("atproto-password")).get()
523
-
model_name = await Variable.get("morning-model", default="claude-sonnet-4-6")
523
+
model_name = await Variable.get("curate-model", default="claude-haiku-4-5")
524
524
print(f"using model: {model_name}")
525
525
526
526
# authenticate