1 / 3
【Overview】 Corrath is a low-latency AI gateway that sits between an application and its model providers. Every completion request passes through a proxy layer that screens the payload, checks it against spending rules, and forwards it to a provider, with the goal of keeping the added latency under 5ms. The problem it solves is the one every team hits after shipping an LLM feature: prompt injections and jailbreak attempts arrive from the outside, token spend can run away inside a billing cycle, and a provider API outage takes the feature down with it. Corrath puts those controls into one gateway layer instead of spreading them across application code, so routing, filtering, budgets, and failover are enforced on the request path itself. It is built for production engineering teams, platform engineers, security directors, and developers building on LLM APIs — people who need one place to set policy on outbound model traffic, see what it costs, and keep the feature online. 【Key features】 • AI Gateway (Proxy Router): Routes completion requests through a proxy layer with high-availability failover, latency logs, and unified streaming completions. • Security Firewall: Detects and blocks prompt injections, jailbreak instructions, adversarial payloads, and PII leaks in real time. • Budget Controls: Enforces granular spending limits per developer key and automatically blocks excess completion calls. • Cost Forecasting: Predicts end-of-month model bills from traffic patterns, flags spend anomalies, and sends Slack warnings before limits are crossed. • Multi-Provider Routing: Routes requests dynamically between OpenAI, Claude, Gemini, Mistral, and DeepSeek using standardized syntax. • Failover Protection: Reroutes completion requests to backup providers when the primary API is down, using automated outage circuit breakers. 【Use cases】 1. Neutralizing prompt injections and jailbreak attempts before they reach a model. 2. Enforcing developer spending limits and catching runaway billing loops. 3. Failing over between AI providers during outages. 【Pricing】 Free: $0/forever — 1 Project Space, 1 Vaulted API Key, 500 Requests / mo, Basic Proxy Gateway. Starter: $0/month — 5 Project Spaces, 10 Vaulted Keys, 50K Requests / mo, Cost Forecasting Engine. Pro: 20 Project Spaces, Unlimited API Keys, 500K Requests / mo, Security Firewall included. Business: Unlimited Projects, Unlimited API Keys, Unlimited Requests / mo, Up to 10 Team Members. Enterprise: CUSTOM — Custom Volume Requests, Unlimited Team Members, On-Premise Support SLAs, Dedicated Operations SLAs. 【Pros and cons】 Pros: - Teams already running LLM traffic in production get injection filtering, budget enforcement, and failover from one layer instead of writing each control into their own application code. - Because the controls sit in the request path, policy applies to every completion call without changing how developers call models. - The Free and Starter tiers cost nothing, so a team can evaluate the proxy, the vault, and cost forecasting before paying for the security tier. Cons: - The Security Firewall is included from the Pro plan; the Free and Starter tiers list the basic proxy gateway and the cost forecasting engine instead. - Spending limits are enforced per developer key, so budgets are scoped to keys rather than to any other grouping. - The Business plan caps team membership at 10 members, so larger teams need the Enterprise plan. 【FAQ】 Q: How much does Corrath cost to start with? A: The Free plan is $0/forever and covers 1 Project Space, 1 Vaulted API Key, and 500 requests per month, while the Starter plan is listed at $0/month with 5 Project Spaces, 10 vaulted keys, and 50K requests per month. Q: Who is Corrath built for? A: It is aimed at production engineering teams, platform engineers, security directors, and developers building on LLM APIs. Q: Which model providers can Corrath route between? A: Requests can be routed dynamically between OpenAI, Claude, Gemini, Mistral, and DeepSeek using standardized syntax, with automatic rerouting to backup providers when a primary API is down.
Free · $0 (1 Project Space, 1 Vaulted API Key, 500 Requests / mo) Starter · $0/month (5 Project Spaces, 10 Vaulted Keys, 50K Requests / mo) Pro (20 Project Spaces, Unlimited API Keys, 500K Requests / mo) Business (Unlimited Projects, Unlimited API Keys, Unlimited Requests / mo) Enterprise · CUSTOM (Custom Volume Requests, Unlimited Team Members, On-Premise Support SLAs)
WPS AI is an AI assistant built into WPS Office that helps users write documents, generate presentation outlines and…
Baidu Comate is an AI coding assistant built by Baidu on its ERNIE model. It goes beyond simple autocomplete: it…
UniConverter is an AI media conversion and enhancement tool for video, audio, and images. It combines everyday…
Added
10/23/2026
Views
633
Clicks
2
6 similar tools to consider instead of Corrath