feat(fireworks): add prompt caching middleware

[Mason Daugherty (mdrxy)]

Fireworks agents can now use FireworksPromptCachingMiddleware to automatically set prompt-cache session affinity from the active thread ID. Cross-provider model fallbacks remove those Fireworks-specific settings before invoking the fallback model.

---

Fireworks prompt caching is most effective when requests from the same agent thread reach the same model replica. Agents currently need to manage that routing affinity themselves, which makes cache reuse easy to miss.

This adds FireworksPromptCachingMiddleware, which derives session affinity from config.configurable.thread_id while respecting affinity settings supplied by the caller. It also updates ModelFallbackMiddleware so Fireworks-only cache settings are retained for Fireworks fallbacks and removed before requests are sent to another provider.

[open-swe]

Open SWE Review found 1 potential issue.

Open in Web • View Open SWE trace

[open-swe]

🔵 Preserve OpenAI prompt cache keys

prompt_cache_key is not Fireworks-specific: ChatOpenAI also exposes it as a supported per-invocation setting for cache routing. Because every non-Fireworks fallback now passes through this branch, an OpenAI primary/fallback request with model_settings={"prompt_cache_key": ...} silently loses the caller's valid key after the primary fails, reducing cache affinity and changing the fallback request. The sanitizer needs to distinguish affinity injected for Fireworks from a provider setting that the selected fallback supports, rather than removing this shared key from every non-Fireworks model.

(Refers to lines 127-129)

---

Your feedback helps Open SWE learn. React with 👍 or 👎 to tell us if this review comment was useful.

[socket-security]

Review the following changes in direct dependencies. Learn more about Socket for GitHub.

Diff	Package	Supply Chain Security	Vulnerability	Quality	Maintenance	License
	langchain@​1.3.13

View full report

[codspeed-hq]

Merging this PR will not alter performance

✅ 2 untouched benchmarks
⏩ 15 skipped benchmarks¹

---

_{Comparing mdrxy/fireworks/prompt-cache-middleware (a2f04da) with master (cdca311)²}

Footnotes

1. 15 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

2. No successful run was found on master (b8a6e36) during the generation of this report, so cdca311 was used instead as the comparison base. There might be some changes unrelated to this pull request in this report. ↩

[open-swe]

Open SWE Review found 1 potential issue.

Open in Web • View Open SWE trace

[open-swe]

🟡 Regenerate the stale Fireworks lockfile

The merge resolution leaves this lockfile out of sync with its editable workspace sources: uv lock --check --directory libs/partners/fireworks now fails, and uv lock --dry-run reports that langchain-core must move from 1.5.0 to 1.5.2 and langchain from 1.3.13 to 1.3.14. Any CI or contributor workflow that checks the package lockfile therefore fails until the Fireworks lockfile is regenerated after the merge.

(Refers to line 819)

---

Your feedback helps Open SWE learn. React with 👍 or 👎 to tell us if this review comment was useful.