Replace expensive managed networking services with a single intelligent gateway. Built-in LLM proxy with response caching. 95% cost savings. Available on AWS, coming soon to GCP and Azure.
Source NAT for internet access, port forwarding for inbound. Zero per-GB fees.
GatewaySync AWS NLB config to IPVS kernel rules. TLS termination via HAProxy.
GatewayRoute to OpenAI, Anthropic, Bedrock via single SDK. Response caching, token counting, cost attribution.
AIPer-AZ ASG with warm pool for sub-10s recovery. EIP failover with per-AZ DNS.
GatewayAuto instance type change with zero downtime. GWLB for multi-instance linear scaling.
GatewaySync rules from AWS Network Firewall -- no endpoints needed ($0/month). ET Open 30K+ threat rules.
SecurityDistributed Squid proxy with per-AZ DNS. NCSA auth, TLS inspection, domain filtering.
SecurityLog every LLM API call with model, tokens, cost, and source IP. Compliance-ready JSON format.
AIAuto-create full-mesh peering across accounts and regions. Free routes, zero data-plane cost.
EnterpriseSubscribe to Camphor AI Gateway on AWS Marketplace. The machine image is automatically configured for your cloud environment.
Choose your VPC (new or existing), instance type, and availability zones. Infrastructure-as-code creates everything automatically.
Point your private subnets to the Camphor gateway. All traffic flows through Camphor for NAT, firewall, and AI proxy.
Replace $500+/month in AWS networking costs with a $30/month t4g.medium. AI proxy included free. Ship faster, spend less.
Route all LLM API calls through a single gateway with caching. Cut API costs 30-70%. Audit every prompt for compliance.
Full-mesh VPC peering across 100+ accounts. IDS/IPS with zero per-GB fees. AI semantic caching at scale.
It's a full EC2 appliance running in your VPC. Traffic never leaves your network. Camphor sits in the data path as a transparent gateway -- your instances route through it like a NAT gateway, but with firewall, load balancing, and AI proxy built in.
Each AZ runs its own Auto Scaling Group with a warm pool for sub-10 second recovery. EIPs automatically fail over. For multi-instance mode, GWLB distributes traffic across multiple gateways with health checking.
Identical prompts return cached responses instantly (0ms latency, zero API cost). The cache is configurable per model and per route. Semantic caching (Enterprise tier) matches similar-but-not-identical prompts using embedding similarity.
Yes. Deploy into an existing VPC or create a new one. The CloudFormation template supports both modes. Just point your route tables to the Camphor ENI and traffic flows through automatically.
Near-zero. NAT and firewall run in kernel space (nftables, IPVS). A c6in.xlarge handles 25+ Gbps. Vertical scaling automatically upgrades the instance type as traffic grows, with zero-downtime migration.
No. Camphor is a network-level appliance. No agents, no sidecars, no code changes to your applications. Just route traffic through the gateway via VPC route tables.
Transit Gateway charges $0.02/GB for every byte. Camphor uses free VPC peering with auto-created routes, full-mesh topology, and cross-account management -- all at $0/GB data transfer cost.
Yes. A custom Terraform provider ships with every release. Deploy via CloudFormation, then configure features (NAT, firewall rules, AI proxy routes, peering) entirely through Terraform or the REST API.
Deploy Camphor AI Gateway in your cloud in under 10 minutes. No agents, no sidecars, no commitments.