TensorWall – Open-source LLM gateway with budget controls and security
TensorWall – Open-source LLM gateway with budget controls and security
Hi HN, I'm the creator of TensorWall. While building LLM-powered applications, I noticed a recurring gap: developers either give their apps a raw API key (risky) or spend weeks building custom proxies to handle rate-limiting, security, and the ever-present fear of exploding API bills. TensorWall is an open-source control plane designed to sit between your applications and your models. It gives you the visibility and guardrails needed for production. Key Features: Unified API: One endpoint for OpenAI, Anthropic, Ollama, and LM Studio Cost & Budget Control: Set hard spending limits and granular rate-limiting per app to prevent "bill shocks" Security: Prompt injection detection (PII redaction on the roadmap) Observability: Full audit logs of every request and token usage (essential for compliance) Deploy in 60 seconds: git clone https://github.com/datallmhub/TensorWall.git cd TensorWall && docker-compose up -d Why Open Source? Security and financial infrastructure shouldn't be a black box. Your AI gateway should be auditable, self-hostable, and community-driven. I'm looking for brutal feedback on the architecture. What are you currently using to keep your LLM costs and security under control? GitHub: https://github.com/datallmhub/TensorWall
Share cardActual performance
Launch Intel predictions
Analyze your own launch →Correct prediction on native model
Similar products
AgentPort – Open-source Security Gateway For Agents
Security gateway for LLM agents
Open Source MCP Gateway
Scarf Gateway is now an open source Haskell project
VernLLM – LLM fallback, no gateway
LLM Gateway and Red Teaming
Your gateway to open source collaboration
Open-Source Data Security Platform
GraphQL Armor – An open source security layer for GraphQL
Apache APISIX is an open-source and dynamic API gateway