Question 1

What if I don't configure any models?

Accepted Answer

Nothing changes. Both tiers start empty — the router does nothing and pi keeps using your default model. Routing only begins once you run /router config and pick your models.

Question 2

Does this actually save money?

Accepted Answer

Yes. Routine turns stay on your fast, value-priced model instead of burning a flagship on every request. The judge itself costs a few thousand tokens at your fast-tier price — the savings dwarf it.

Question 3

Is it safe to try?

Accepted Answer

Completely. The router does nothing until you configure it, it is pure TypeScript with zero runtime dependencies, and /router off disables it instantly. No lock-in.

Question 4

What exactly does the smart tier do?

Accepted Answer

It's the CTO role — it sets direction, corrects course, reviews results, and takes on the hard problems itself: architecture, design review, security, multi-step plans, irreversible changes. High-stakes turns don't get dropped. It doesn't just advise or review; it writes the code and does the work.

Question 5

Does the Judge add noticeable latency?

Accepted Answer

No. The classification call is a few thousand tokens and takes about 200ms–2s. You'll see 🧭 judging… in the status bar during the call — most users never notice it.

Question 6

What if my primary model 429s or times out?

Accepted Answer

You keep working. The failed model enters an exponential-backoff cooldown — 5xx starts at 1m (1m → 4m → 16m → 1h → 4h… capped at 6h), while a failover-worthy 4xx (429 rate limit / quota) skips the first tiers and starts at 16m, because client-side limits usually outlive server blips. The next healthy model in the same tier takes over automatically — even mid-turn. A successful response clears the cooldown.

Question 7

What is task-level orchestration?

Accepted Answer

When the Judge says a task is complex, the Smart tier runs as a CTO: it plans the work, delegates implementation to Fast engineer subagents (fresh context, model pinned from your Fast tier), reviews each result, iterates with concrete feedback, and does a final acceptance pass. Simple tasks never orchestrate — they stay on the plain router. Orchestration is on by default (auto mode); /router orchestrate off disables it. It needs the pi-subagents extension — without it, complex tasks run directly on the Smart tier, exactly as before.

Question 8

Does orchestration cost more?

Accepted Answer

No — that's the point. The Smart tier plans and reviews; Fast subagents implement, each in a fresh context that's small and cheap: ~$0.004 for a narrow task vs ~$0.06 for an inherited 176k-token fork. Hard caps keep it bounded: maxRounds 3 and escalationThreshold 2. If a task turns out simple, the machinery never engages.

Question 9

What keeps an orchestrated turn honest?

Accepted Answer

Two safety nets. The convergence protocol requires every re-delegation to carry a structured failure report — what failed, where, the exact acceptance test to re-run — and re-sending identical feedback is forbidden (the CTO takes the phase over). On top of that, a post-turn acceptance audit (v1.3.0) runs after every orchestrated turn: deterministic checks always run (every worker reported back, the final message carries a CTO summary, whether a hard cap was hit), plus an optional small fast-tier LLM audit that checks grounding, goal alignment, and delivered quality against your original request. Findings never block the finished turn — they surface as console.warn + toast and in /router status as "Last audit". Toggle with orchestration.audit.enabled (default on).

Question 10

Can I force a specific model for one turn?

Accepted Answer

Yes. /route-force pins Smart or Fast for the next turn; /route-force / pins an exact model. /route-force auto clears the override.

Question 11

Does it work with models from different providers?

Accepted Answer

Yes — mix freely. Each tier is an ordered list of {provider, model, priority} pairs, so your fast tier can be DeepSeek and your smart tier Kimi, all from one config.

Question 12

Can I monitor how it's performing?

Accepted Answer

Yes — run /router stats to see per-tier spend and savings, window size and confidence distribution (high / mid / low / none), cumulative upgrade and downgrade counts, and tokens-per-second. It reports cost telemetry: how much each tier spent and a baseline — what the session would have cost if every turn ran on your Smart model (i.e. no router). Spend: fast $0.045 (9 calls) · smart $0.42 (3 calls) · total $0.465, baseline $3.21, saved $2.74. The status bar also shows live throughput: [🧠 kimi-k3 • 23 tok/s].

Question 13

How do I tune the downgrade behavior?

Accepted Answer

Two thresholds: `threshold` (default 0.6) controls how many of the last 5 classified turns must weigh toward fast before a downgrade happens; `minConfidence` (default 0.5) sets the floor below which votes are ignored entirely. Raise `threshold` to 0.8 to stay on Smart longer. Upgrades are always instant — only downgrades wait.

Question 14

How is it different from pi-bifrost and pi-smart-router?

Accepted Answer

Both are pi routers with the same goal. @tenchi4u/pi-bifrost decides with 7-step rules + history tricks across 4 tiers — more cases, but harder to reason about; it saves subscription quota rather than dollars, and its circuit breaker can switch tiers on failure. pi-smart-router runs a 12-step local pipeline (no LLM) with 3 tiers including a local one, estimates cost with a formula, and needs a local DB + ML model (~2.5 MB plus downloads) and 15+ env vars. Shift Router keeps the decision to one readable LLM prompt (auditable plain text, JSON-mode enforced), two tiers with one mental model, orchestration (the Smart CTO delegates to Fast subagents), dollars-saved cost telemetry, same-tier failover with a cooldown shared with the judge, and ~409 KB with zero deps. CC-Switch is a different category entirely: a desktop app for manually switching provider configs between sessions — it coexists fine.

Question 15

Is there a version for DeepSeek Harness?

Accepted Answer

Yes — dsh-shift-router (v0.5.0), a DSH adaptation of the same router. It plugs into the harness via cordis.patch.yml, routes through the harness's own agent/request pipeline, and configures live two ways: a full GUI card (Settings → Plugins → Plugin configuration — the "Shift-Router" card, with both tier chains editable) and an interactive /router config editor (get/set/unset/diff). Same capabilities as pi: LLM Judge, cache-aware routing, runtime failover, task-level orchestration, cost telemetry — with orchestration caps enforced by the plugin, not just prompted. 109 unit tests plus a credential-free headless e2e. Install: git clone https://github.com/green-dalii/dsh-shift-router && npm install && npm run build, then dsh plugin --profile web add <path>.

Question 16

On DeepSeek Harness, are subagents routed too?

Accepted Answer

No. Subagents spawned by the subagent tool carry origin === 'subagent' and keep their pinned model — the router only drives top-level agents. During orchestration, the CTO delegates to Fast workers whose model is pinned by deployment configuration (tool-subagent agentOptions), so workers always run at the Fast tier's price.

Setup	Fast 🦾	Smart 🧠
≤ 32 GB	LFM2.5-8B-A1B (8.5B), granite-4.1-8b (8.8B), Qwen3.6-27B (dense, q4 ≈ 14 GB), gemma-4-26b-a4b-it (q4 ≈ 13 GB), Qwen3.6-35B-A3B (MoE 36B/3B, q4 ≈ 18 GB), Laguna-XS-2.1 (MoE 33B/3B, q4 ≈ 17 GB), Ternary-Bonsai-27B (1.58-bit ≈ 7 GB, laptop/phone)	Cloud frontier model
32–128 GB	Fable-Fusion-711-NEO-MAX-MTP GGUF (27B, 2026-07, best post-trained), gemma-4-31B-it (q4 ≈ 16 GB), Laguna-S-2.1 (117.6B, q4 ≈ 59 GB — needs 64 GB+)	Fable-Fusion-711 base weights, or cloud frontier
≥ 128 GB	Fable-Fusion-711-NEO-MAX-MTP GGUF (same pick — post-training beats raw size)	DeepSeek-V4-Flash (284 B MoE / 13 B active, UD-Q4_K_XL ≈ 155 GB — needs 192 GB+ unified memory; on 128 GB-class use 1–2 bit ternary)

Setup	Fast 🦾	Smart 🧠
Anthropic	sonnet-5	opus-5 / fable-5
OpenAI	gpt-5.6-luna	gpt-5.6-sol
Google	gemini-3.5-flash-lite	gemini-3.6-flash
Qwen (AliBaba)	qwen3.7-plus	qwen3.8-max
DeepSeek	deepseek-v4-flash	deepseek-v4-pro
Z.AI (GLM)	glm-5 / glm-5-turbo	glm-5.2
xAI (Grok)	grok-4.5-fast	grok-4.5

Setup	Fast 🦾	Smart 🧠
DeepSeek Harness (DSH)	deepseek-v4-flash	deepseek-v4-pro
Lowest cost	deepseek-v4-flash	claude-opus-5
Multi-provider fallback	deepseek-v4-flash + glm-5.2	claude-opus-5 + gpt-5.6-sol + kimi-k3
Flat-fee	opencode-go/deepseek-v4-flash	opencode-go/glm-5.2
1 M context, long repo	deepseek-v4-flash	gemini-3.6-pro / kimi-k3
Multimodal	deepseek-v4-flash	claude-opus-5 (vision) / gemini-3.6-pro
Multilingual / Chinese-first	deepseek-v4-flash	qwen3.8-max
GDPR / Europe	deepseek-v4-flash (via OpenRouter)	mistral-medium-2604

Pick your model pair

Coding subscription plans

Local models by memory

Same-provider tier ladder

Cross-provider pairing