Claude Model Router

A Claude Code Mod that chooses a model and effort for each main-conversation turn, advised by a prompt classifier: Jev; Cloudflare's Clef, Clef Flash or Clef Omni; OpenAI; or a local Ollama model.

The Mod changes only the model and effort in Claude Code's turn hook. Claude Code sends the request, streams the reply, runs tools, and reports usage. The router does not proxy model traffic or start a local server; an Ollama classifier calls an Ollama server that you run. Subagents get one model each when they start, chosen the same way: every agent whose definition names no model, says model: inherit, or sets modelRouting: auto. Their steps are never re-routed. See Subagent routing.

Providers

Use the provider that already works with Claude Code. Select its router profile, not a new endpoint:

CLAUDE_ROUTER_PROVIDER=deepseek claude --model 'deepseek-flash[1m]'

Create ~/.claude/deepseek-router.json containing {} to use DeepSeek defaults. With CLAUDE_CONFIG_DIR, put the file in that directory instead.

Built-in profile	Router file	Default models
Anthropic, the default	router.json	Haiku 5.5, Sonnet 5.5, Opus 5.5; Fable 5.2 for routes you set
deepseek	deepseek-router.json	DeepSeek Flash, routed by effort
zai	zai-router.json	GLM-5.3-Flash and GLM-5.3
minimax	minimax-router.json	MiniMax-M3.1-Flash, routed by effort
mimo	mimo-router.json	MiMo-V2.6-Flash and MiMo-V2.6-Pro, no routed effort
kimi	kimi-router.json	Kimi-K3.1-Pro, routed by effort

OpenRouter, Ollama, other local servers and other compatible hosted providers use custom <provider>-router.json files. They have no built-in model defaults. A selected file must exist and validate, or the router leaves Claude Code's model unchanged.

See Providers for default routes and activities, custom profiles, and the --settings flag alternative. Execution provider selection is separate from classifier selection. Observed usage stays visible even when a provider's cache behavior makes savings comparisons unavailable.

Supported classifiers

The active classifier is asked at the start of each logical turn, for the tier, task goal and current phase, and again in the background every 4 requests while the turn runs (policy.reviewEveryRequests). Pick one on the pane's Classifier tab or in the selected router file; only the active one receives prompt text. With subagent routing on (the default), it is also asked once per routed subagent, with the task's description and prompt; set "subagentRouting": "off" to keep that text local.

Classifier	Service	Credential	Prompt text goes to	How it is asked for the activity
Jev	typesafe.ai	Jev API key	typesafe.ai	Goal and phase choice questions in the same request
Clef	Cloudflare Workers AI, 27B	Cloudflare API token and account ID	Cloudflare	The same
Clef Flash	Cloudflare Workers AI, 9B	Cloudflare API token and account ID	Cloudflare	The same
Clef Omni	Cloudflare Workers AI, 30B MoE	Cloudflare API token and account ID	Cloudflare	The same
OpenAI	OpenAI, gpt-6-luna	OpenAI API key	OpenAI (api.openai.com)	Goal and phase choices in questions, in the same request
Ollama	A local Ollama model	None	Stays on this machine	Goal only in a second request; no phase or continuation answer

Jev 1.14 is the default; a configured model override stays in force. Configuration gives the endpoints, timeouts, and how to add another service that speaks one of the supported protocols.

The band above the prompt shows the tier, activity, model, and reason for each turn. /router opens a pane to pin a tier, edit the model and effort of each tier, and tune the policy. See the user guide.

The router needs Claude Code 2.1.289 or newer. Savings are not measured. The pane's Usage tab estimates what routing cost against no routing (your own model on the main thread, Claude Code's pick for each subagent) for the same tokens at configured list prices, for this session, today, 7 or 30 days, this month or all; it is not a bill and does not judge answer quality. See the evaluation.

How it works

sequenceDiagram
  actor Dev as Developer
  participant Code as Claude Code
  participant Mod as Router Mod
  participant Cls as Classifier
  participant API as Model provider
  Dev->>Code: Prompt
  opt Subagent spawn
    Code->>Mod: agent.spawn
    Mod-->>Code: One model for the agent (on), or none (off)
  end
  Code->>Mod: Main or subagent step
  alt Main conversation, routing on
    Mod->>Cls: Bounded prompt and dialogue
    Cls-->>Mod: Tier advice
    Mod-->>Code: Selected model and effort
  else Routing off or subagent step
    Mod-->>Code: Original model and effort
  end
  Code->>API: Native request and tools
  API-->>Code: Response stream and usage
  Mod-->>Dev: Status band and router pane

The active classifier labels each new logical turn, and the router checks the route again while the task runs: a turn labelled ops that edits code moves up to the code route at once, and a background classifier review every 4 requests can move the tier or the activity either way. A cheaper route on the same task waits for a dwell and a checkpoint its work can show: a passed test, build or lint run, or for commands, docs and read-only work the success the router can see. A notification is treated as progress, but cannot justify a downgrade because it may belong to an older task. /router decisions lists why each turn ran where it did. See Changes during a task. A local policy applies vote, context, failure, and cache-cost rules before the Mod changes the next step. Claude Code owns request construction, model credentials, tools, streaming, and the API cost ledger.

With the Anthropic profile, the base routes are:

Tier	Default model	Effort
micro	Haiku 5.5	medium
low	Haiku 5.5	high
medium	Opus 5.5	medium
high	Opus 5.5	xhigh

These are routing defaults. They do not claim equal model quality. Configuration gives the published results behind them and the router.json lines that restore the 1.3 Sonnet ladder.

Activity routing classifies the task's goal (code, debug, explore, plan, review, ops, or docs) separately from the current phase. Bounded verification or execution can use the ops route without relabeling the goal. A route is then picked by tier and effective activity. It is on by default since 1.7. A · cell uses the tier's own route, the base route:

The Anthropic profile uses these activity defaults. Other profiles have their own defaults.

Activity	micro	low	medium	high
Base route	Haiku · medium	Haiku · high	Opus · medium	Opus · xhigh
code, debug, plan, review	·	Sonnet · high	·	·
ops	·	·	Haiku · high	·
explore, docs	·	·	·	·

/router activities off stops asking and keeps the base routes. Activity routing explains each cell and how to change it. The label is optional: on uncertain or a weak label the base route is used if the move passes the cache check, when the activity answer fails the running activity stays and the tier answer still applies, and when the classifier fails the running route stays. See Route by activity, and Evaluation for how it is checked. Its default overrides rest on published vendor results and a list-price replay of one developer's sessions, not measured savings.

A routing mode trades cost against answer quality without new routes: cost takes a cheaper model or one effort level less where a turn allows it, quality one effort level more or the tier above, and balanced, the default, runs the routes as configured. Set it per session with /router mode cost, per project in .claude/router.json, or for the profile with "mode" in router.json.

For mixed Anthropic and DeepSeek sessions, use the explicit model-proxy profile with a separately started local proxy. No built-in mixed-provider defaults are installed.

Install

You need Claude Code 2.1.289 or newer and credentials for one supported classifier.

claude plugin marketplace add alexei-led/claude-router
claude plugin install router@alexei-led-claude-router

For the default Anthropic profile, start on the full baseline model, for example claude --model claude-haiku-5-5. For other providers, select a provider profile. Run /plugin configure router and save the key. For any other classifier, save its credentials, or run ollama serve with a pulled model, then pick the classifier on the pane's Classifier tab. A new session on a model that a tier routes to, or with activity routing on a model an activity override uses (Sonnet 5.5 on the defaults), starts with routing on; on any other model it starts with routing off and says why, and /router auto turns routing on.

Claude Code updates the plugin at startup when auto-update is on for this marketplace. Otherwise run claude plugin marketplace update alexei-led-claude-router, then claude plugin update router@alexei-led-claude-router.

Upgrading from 0.8? Remove the gateway settings that 0.8 /router:setup wrote before the first 1.0 session, or requests go to a gateway that 1.0 no longer starts. Follow the migration steps.

Run from a checkout

Start Claude Code with this directory as a local plugin and the full Haiku baseline model:

claude --plugin-dir "$PWD" --model claude-haiku-5-5

In Claude Code, run /plugin configure router and save your classifier key in the sensitive plugin option. The Mod adds its status band and /router pane. No status-line setup or gateway environment variables are needed.

Run /router to open the pane: route status, per-tier controls, tuning, and usage, including routing vs no routing. Run /model to select a model and turn routing off, and /router auto to turn it back on. For controls, metrics, tuning, and troubleshooting, see the user guide.

Before a push, run the same checks as CI. npm run setup installs Biome and TypeScript into tools/; the plugin root keeps no lockfile, so Claude Code installs nothing with the plugin. npm run validate and npm run test:plugin need the claude CLI: they load the Mod in the Claude Code engine, which refuses some faults that lint and unit tests miss.

npm run setup
npm run check && npm run typecheck && npm test && npm run validate && npm run test:plugin

Documentation

• User guide: install, use the controls, read the panel, and troubleshoot.
• Providers: provider names, router files, built-in defaults, and custom profiles.
• Configuration: router-file settings, classifier keys, cache policies, and migrations.
• Architecture: Mod event flow, state, safeguards, and runtime boundaries.
• Activity routing: the default matrix by activity, why each cell, and how to change or turn it off.
• Routing modes: cost, balanced and quality, where the mode comes from, and every decision per mode.
• Native router details: panel semantics, tuning, and accepted limits.
• Evaluation: historical gateway results and the native shadow replay.
• Changelog: user-visible changes by version.