tokenRoute
Eval gate
AI routing intelligence

Model routing, backed by benchmark evidence.

TokenRoute compares your real prompt workloads across local, private, and BYOK cloud models, then exports the route decision your apps can use.

Benchmark route story
prompt to route decision
live flow
prompt: I was charged twice after upgrading my plan yesterday.
promptcloudlocalprivatejudge 1judge 2policy
1 policy
Recommended for this benchmark

The result includes recommendation, fallback order, score evidence, latency, and confidence.

route_decision:
{
  model: "qwen-2.5-7b",
  score: 94, latency: "1.28s",
  fallback: ["llama3.2"]
}
Prompt in
Contract captured
Route
Candidates run
Judge
Evidence scored
Result
Policy exported
Benchmark
Prompt contract against selected models
Score
Quality, latency, cost, and failure evidence
Route
API decision and gateway-ready policy
What TokenRoute is

A benchmark intelligence layer for AI model selection.

Not another generic gateway. TokenRoute decides what your gateway, app, or agent should route to.

Benchmark packs

Repeatable runs across candidate models with score, latency, cost, failures, and raw output evidence.

Route-decision API

Apps and agents ask which model to use for a workload under quality, latency, and cost constraints.

Policy exports

Push advisory fallback order and thresholds into gateways such as LiteLLM or your own router.

Benchmark intelligence

Track model/provider trends by workload category without exposing raw prompts across tenants.

Integration-first

Bring your models. Keep your stack.

TokenRoute is the glue layer between model catalogs, local connectors, gateways, CI, and production apps.

Connector surface
available and planned integration paths
Local/private
Ollama and private endpoints through the Local Connector Agent
BYOK cloud
OpenRouter, OpenAI, Anthropic, and provider adapters
Gateways
LiteLLM-style policy export first, live routing later
Engineering
JSON/CSV packs, CI gates, and route-decision API
Why it matters

Make model changes defensible.

No blind routing

Every recommendation links back to benchmark evidence.

BYOK friendly

Use local, private, and cloud models without TokenRoute carrying inference spend.

Private intelligence

Aggregate trends without exposing raw prompts across tenants.

Pricing

Adoption-first pricing for model-routing teams.

Bring your own provider keys. TokenRoute charges for benchmark intelligence, route-decision usage, and evidence history, not token resale.

Free

$0
per month

Useful trial for validating a real workload before committing.

Included
10 benchmark packs / month
1,000 route API calls / month
1 project and 1 user
30-day history
3 models per run
25 prompts per run
BYOK and local/private models
Start free

Builder

Recommended
$9
per month

Low-friction plan for builders, small apps, and client experiments.

Included
75 benchmark packs / month
10,000 route API calls / month
3 projects and 2 users
90-day history
5 models per run
50 prompts per run
Exports, cost caps, and provider health
Choose Builder

Team

$29
per month

Shared benchmark evidence for early AI product and platform teams.

Included
300 benchmark packs / month
50,000 route API calls / month
10 projects and 5 users
1-year history
Shared benchmark packs
Route-decision API
Launch-gate reports and priority support
Choose Team
Overage proposal: $5 per 100 extra benchmark packs.
Route API overage proposal: $2 per 100,000 extra calls.
Enterprise remains custom for SSO, audit, retention, private deployment, and procurement.
Ready for the build path.

Start with a benchmark pack, inspect the evidence, then export an advisory route policy.