Ask HN: How do you keep 54 LLM workflows on the right models?
I have a Django app with 54 LLM-backed workflows. Up until recently I've exclusively used Anthropic models via AWS Bedrock but just set up OpenRouter to test the new Gemini models given they seem to match Sonnet/Haiku intelligence but with 3-5x output speed.
- ▪I have a Django app with 54 LLM-backed workflows.
- ▪Up until recently I've exclusively used Anthropic models via AWS Bedrock but just set up OpenRouter to test the new Gemini models given they seem to match Sonnet/Haiku intelligence but with 3-5x output speed.
Hacker News (AI / LLM) files mainly under ai. We currently carry 4,500 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | Ycombinator |
| Canonical URL | https://news.ycombinator.com/item?id=49262312 |
| Publication time | Tue, 11 Aug 2026 18:17:16 +0000 |
| Retrieval time | 2026-08-11T18:20:42.106Z |
| Last seen | 2026-08-11T18:20:42.106Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | qWNc_xCodNiX · 1 stories |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
I have a Django app with 54 LLM-backed workflows. Up until recently I've exclusively used Anthropic models via AWS Bedrock but just set up OpenRouter to test the new Gemini models given they seem to match Sonnet/Haiku intelligence but with 3-5x output speed. I'm using Pydantic AI for validation/normalization.I currently maintain a registry that describes a workflow's purpose, what we're optimizing for (intelligence, speed, cost), its eval, and a human-readable bar that must be achieved.I'm curious what strategies/systems people are using to keep track of everything and ensure an optimal model is being used for a given workflow.
Excerpt limited to ~120 words for fair-use compliance. The full article is at Ycombinator.