Model failover
Fased handles failures in two stages:- Auth profile rotation within the current provider.
- Model fallback to the next model in
agents.defaults.model.fallbacks.
Auth storage (keys + OAuth)
Fased uses auth profiles for API keys, provider tokens, and OAuth credentials.- Secrets live in
~/.fased/agents/<agentId>/agent/auth-profiles.json(legacy:~/.fased/agent/auth-profiles.json). - Config
auth.profiles/auth.orderare metadata + routing only (no secrets). - Legacy import-only OAuth file:
~/.fased/credentials/oauth.json(imported intoauth-profiles.jsonon first use).
type: "api_key"->{ provider, key }type: "token"->{ provider, token, expires? }type: "oauth"->{ provider, access, refresh, expires, email? }(+projectId/enterpriseUrlfor some providers)
Profile IDs
OAuth logins create distinct profiles so multiple accounts can coexist.- Default:
provider:defaultwhen no email is available. - OAuth with email:
provider:<email>(for exampleanthropic:[email protected]).
~/.fased/agents/<agentId>/agent/auth-profiles.json under profiles.
Rotation order
When a provider has multiple profiles, Fased chooses an order like this:- Explicit config:
auth.order[provider](if set). - Configured profiles:
auth.profilesfiltered by provider. - Stored profiles: entries in
auth-profiles.jsonfor the provider.
- Primary key: profile type (OAuth, then token, then API key).
- Secondary key:
usageStats.lastUsed(oldest first, within each type). - Cooldown/disabled profiles are moved to the end, ordered by soonest expiry.
Session stickiness (cache-friendly)
Fased pins the chosen auth profile per session to keep provider caches warm. It does not rotate on every request. The pinned profile is reused until:- the session is reset (
/new//reset) - a compaction completes (compaction count increments)
- the profile is in cooldown/disabled
/model ...@<profileId> sets a user override for that session
and is not auto-rotated until a new session starts.
Auto-pinned profiles (selected by the session router) are treated as a preference:
they are tried first, but Fased may rotate to another profile on rate limits or timeouts.
User-pinned profiles stay locked to that profile; if it fails and model fallbacks
are configured, Fased moves to the next model instead of switching profiles.
Why OAuth can “look lost”
If you have both an OAuth profile and an API key profile for the same provider, round-robin can switch between them across messages unless pinned. To force a single profile:- Pin with
auth.order[provider] = ["provider:profileId"], or - Use a per-session override via
/model ...with a profile override (when supported by your UI/chat surface).
Cooldowns
When a profile fails due to auth, rate-limit, overload, model-not-found, or session-expired errors, Fased records that profile in cooldown and moves to the next usable profile. Timeout and provider format errors can still trigger profile rotation or model fallback for the current run, but Fased does not mark the auth profile in cooldown for those reasons. A timeout is model/network-specific, and a format error is usually request-shape specific; marking the credential would block other valid model routes on the same provider. OpenRouter profiles intentionally bypass auth-profile cooldown tracking because OpenRouter is itself a multi-provider router. Fallback and runtime errors still surface normally. Cooldowns use exponential backoff:- 1 minute
- 5 minutes
- 25 minutes
- 1 hour (cap)
auth-profiles.json under usageStats:
Billing disables
Billing or credit failures are treated as failover-worthy, but they are usually not transient. Instead of a short cooldown, Fased marks the profile as disabled with a longer backoff and rotates to the next profile or provider. State is stored inauth-profiles.json:
- Billing backoff starts at 5 hours, doubles per billing failure, and caps at 24 hours.
- Backoff counters reset if the profile hasn’t failed for 24 hours (configurable).
Model fallback
If all profiles for a provider fail, Fased moves to the next model inagents.defaults.model.fallbacks. This applies to failover-worthy errors such
as auth failures, rate limits, billing disables, overload, model-not-found,
format errors, and timeouts after the current route has no usable profile left.
Other application errors do not advance fallback.
When a run starts with a model override (hooks or CLI), fallbacks still end at
agents.defaults.model.primary after trying any configured fallbacks.
Related config
See Gateway configuration for:auth.profiles/auth.orderauth.cooldowns.billingBackoffHours/auth.cooldowns.billingBackoffHoursByProviderauth.cooldowns.billingMaxHours/auth.cooldowns.failureWindowHoursagents.defaults.model.primary/agents.defaults.model.fallbacksagents.defaults.imageModelrouting