Cloudflare Opens Its Internal AI Stack to Everyone
Cloudflare's CTO says the company opened up the internal AI stack it runs on: AI Gateway and MCP portals, now available in every customer's dashboard. One endpoint fronts every model provider, including OpenAI, Anthropic, Google, Kimi, GLM, and Cloudflare's own Workers AI. Spend shows up in one view, with limits by model, provider, team, user, or app. Keys stay behind the gateway, and Data Loss Prevention scans every prompt, response, and tool call. MCP portals decide which servers and tools each agent can use, and access can be revoked instantly. Code Mode turns an agent's MCP tools into an API, which Cloudflare says cut its own token costs by about 93%.
Transcript
Cloudflare's CTO faced forgotten keys and unexplained bills, so the company built controls and shared them.
Inside Cloudflare, six providers supplied models. Keys got pasted into scripts, bills went unexplained, nobody was watching.
Every AI call and agent now runs through two dashboard products, AI Gateway and MCP portals.
One endpoint fronts every provider, including OpenAI, Anthropic, Google, and Workers AI, on one bill.
Spend shows in one view, with caps by model or team. Data Loss Prevention scans every prompt and tool call.
MCP portals grant each agent explicit access, revocable anytime. Code Mode turns tools into an API, cutting token costs ninety three percent.
Inboxsmith gives your business a smart AI receptionist. Please like and subscribe for more news.
Sources
Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:
- Cloudflare's CTO faced messy AI usage with forgotten API keys and token bills nobody could explain.(Cloudflare - official announcement)
- Every AI call and agent at Cloudflare now runs through two products in the dashboard, AI Gateway and MCP portals.(Cloudflare - official announcement)
- A single endpoint fronts every provider including OpenAI, Anthropic, Google, Kimi, GLM, and Cloudflare's own Workers AI, with one bill or your own keys.(Cloudflare - official announcement)
- Every request, token, and dollar shows in one view, with limits set by model, provider, team, user, or app.(Cloudflare - official announcement)
- Data Loss Prevention scans every prompt, response, and tool call so confidential data does not leak.(Cloudflare - official announcement)
- MCP portals decide which servers and tools each agent can use, granted explicitly and revocable.(Cloudflare - official announcement)
- Code Mode turns an agent's MCP tools into an API and cut token costs by about ninety three percent.(Cloudflare - official announcement)