Source of truth: , Last updated: 2026-05-13 — v3.8.0

produced by thinking-mode models and replays it transparently on multi-turn requests when the upstream provider requires it. This eliminates the HTTP 400 errors that strict providers raise when a client's conversation history is missing the prior turn's reasoning.

previous assistant message includes the original . The upstream returns 400 with messages like:

from the history they replay. OmniRoute restores it from a server-side cache so the request the upstream sees is consistent. Issue #1628 introduced the hybrid memory/SQLite persistence so the cache survives process restarts.

(two sites, around lines 4093 and 4380). Replay happens in after schema coercion but before dispatch.

(LRU-by-creation) backed by a SQLite table for crash recovery and dashboard visibility.

	in
	table ()

Defaults:

()
• ()
• first

CREATE TABLE IF NOT EXISTS reasoning_cache (
  tool_call_id   TEXT PRIMARY KEY,
  provider       TEXT NOT NULL,
  model          TEXT NOT NULL,
  reasoning      TEXT NOT NULL,
  char_count     INTEGER NOT NULL DEFAULT 0,
  created_at     TEXT NOT NULL DEFAULT (datetime('now')),
  expires_at     INTEGER NOT NULL
);

, , , . is stored as Unix epoch seconds; the SELECT layer normalizes legacy text values via .

returns . The function checks two lists in .

Provider IDs (exact match, case-insensitive):

Model regex patterns (case-insensitive):

. Both require management authentication ( from ).

		clamped to )

GET response shape:

{
  "stats": {
    "memoryEntries": 12,
    "dbEntries": 47,
    "totalEntries": 47,
    "totalChars": 138291,
    "hits": 84,
    "misses": 6,
    "replays": 81,
    "replayRate": "90.0%",
    "byProvider": { "deepseek": { "entries": 32, "chars": 98412 } },
    "byModel": { "deepseek-reasoner": { "entries": 32, "chars": 98412 } },
    "oldestEntry": "2026-05-13T10:00:00.000Z",
    "newestEntry": "2026-05-13T11:42:11.000Z"
  },
  "entries": [
    {
      "toolCallId": "call_abc",
      "provider": "deepseek",
      "model": "deepseek-reasoner",
      "reasoning": "...",
      "charCount": 3128,
      "createdAt": "...",
      "expiresAt": "..."
    }
  ]
}

Cleanup: purges expired memory entries and runs . Health-check workers call this periodically.
• Crash recovery: After a restart, memory is empty but the DB still holds unexpired entries. The first lookup for a given is a DB hit; subsequent lookups are memory hits.
• No reasoning, no cache: returns when the assistant message has no / field, so non-thinking responses cost nothing.
• Non-strict providers: When is and the target format is OpenAI, the translator strips any field from outgoing messages — OpenAI Chat Completions does not accept it.

RESILIENCE_GUIDE.md — circuit breakers, cooldowns, model lockouts
• TROUBLESHOOTING.md — diagnosing upstream 400s
• , ,