NEWAWS Bedrock routing + live rate matrix

The AI gateway forreal-time cost visibility

TokenValves sits between your stack and every LLM provider. Track token spend to the cent, enforce hard budget caps, block prompt injection, and route traffic across OpenAI, Anthropic, Gemini and Bedrock with one line of code.

Start Free Proxy Live Rates Matrix

No credit card. 10M governed tokens free every month.

request pathp50 overhead 3.2ms

App Stack

Web & Mobile Apps

AI Agents

Internal Tools

TokenValves Gateway

gateway.tokenvalves.io/v1

Spend caps

Injection shield

PII masking

Fallback routing

Providers

OpenAI

Anthropic

Google Vertex

AWS Bedrock

<5ms

Proxy overhead

4.2B

Tokens governed / day

38%

Avg. spend reduction

99.99%

Gateway uptime SLA

Live rates matrix

Every model. Every rate. Normalized.

Compare input, cached-input and output pricing across providers, refreshed automatically from live provider catalogs with an embedded reference fallback.

Loading reference matrix

Provider / Model	Tier	Input / 1M	Cached / 1M	Output / 1M	Caching
OpenAIGPT-4o / GPT-5 class	Flagship	$2.50	$1.25	$15.00	~50% - 90% off cached inputs
OpenAIGPT-4o mini	Mini / Lightweight	$0.15	$0.075	$0.60	~50% off cached inputs
OpenAIo1 / o3	Reasoning / Frontier	$15.00	$7.50	$60.00	~50% off cached inputs
AnthropicClaude Opus	Frontier	$15.00	$1.50	$75.00	Up to 90% off read cache
AnthropicClaude Sonnet	Balanced	$3.00	$0.30	$15.00	Up to 90% off read cache
AnthropicClaude Haiku	Fast	$0.80	$0.08	$4.00	Up to 90% off read cache
GoogleGemini Pro	Flagship	$1.25 - $2.00	$0.625 - $1.00	$10.00 - $12.00	50% discount on batching
GoogleGemini Flash	Fast / Lightweight	$0.10 - $0.30	$0.05 - $0.15	$0.40 - $2.50	Generous free tier available

Cached = input rate after the maximum provider cache or batch discount. USD, list price.Auto-refreshes every 5 min

Guardrails playground

Flip a switch. Watch the gateway react.

This simulated traffic stream runs through the same policy engine shipped in production. Toggle guardrails and see verdicts, cost and latency change in real time.

PII MaskingRedact emails, SSNs, cards and phone numbers before they leave your network.Prompt Injection BlockingScore every prompt for jailbreak and exfiltration patterns. Block above 0.90.Hard Spend Cap: $5,000/moReturn 429 once month-to-date spend reaches the org budget.

Month-to-date spend$4,986.40 / $5,000

Simulation starts close to the cap so you can see it trigger.

tokenvalves tail --live

streaming

Tokens intercepted

Spend this session

$0.00

Avg overhead

0.0ms

Blocked

Waiting for traffic... press Resume to start the stream.

Simulated traffic. No prompts leave your browser.

Integration

Ship governance in one line.

TokenValves speaks the OpenAI-compatible protocol natively and translates to Anthropic, Gemini and Bedrock behind the scenes. Keep your SDK, change the base URL.

01Create a gateway key

Scope keys per team, environment or customer with their own budgets.

02Swap one line

Point your existing SDK at the TokenValves base URL. No new client library.

03Observe instantly

Tokens, cost, latency and policy verdicts stream to your dashboard.

app.ts+ 2 lines changed

1import OpenAI from "openai";
2 
3const client = new OpenAI({
4  apiKey: process.env.OPENAI_API_KEY,
5  baseURL: "https://gateway.tokenvalves.io/v1",
6  defaultHeaders: { "x-tv-key": process.env.TOKENVALVES_KEY },
7});
8 
9const res = await client.chat.completions.create({
10  model: "gpt-4o",
11  messages: [{ role: "user", content: "Summarize Q3 churn" }],
12});

ROI calculator

See what you're leaving on the table.

Model your savings from cache-aware routing, automatic fallback to cheaper tiers, and hard caps on runaway agent loops.

Monthly token spend$40,000

$1K$500K

Expected prompt caching35%

0%90%

Smart fallback routingRoute eligible traffic to cheaper tiers when quality allows.

Assumes 62% of spend is input tokens, a 75% blended cache discount, and 14% reduction from routing. Results are estimates.

Estimated monthly savings

12,799

-32.0% spend153,583 per year

Prompt caching6,510

Fallback routing4,689

Runaway loop caps1,600

Start saving with the free proxy

Enterprise trust

Compliant proxying, by default.

Your security team gets the controls they need. Your developers keep shipping.

SOC 2 Type II

Annual audit across security, availability and confidentiality.

HIPAA

BAA available. PHI redacted before it reaches any provider.

GDPR

EU data residency with regional gateway pinning.

Zero retention

Prompts are never stored unless you opt in to log capture.

SSO + SCIM

Okta, Entra ID and Google Workspace with role-based policies.

Self-host / VPC

Deploy the gateway inside your own cloud with Helm or Terraform.

Put a valve on every token.

Free up to 10M governed tokens per month. Enterprise plans with VPC deploy and custom SLAs.

Start Free Proxy Talk to Sales