|
Cloudflare TV

Cloudflare Connect 2026

Catch up on everything from Connect 2026.

Watch keynotes and sessions on shipping agentic apps, governing AI traffic, and unifying security.

Check out the Cloudflare Connect Hub to follow along on the conference.

Watch keynotes and sessions on shipping agentic apps, governing AI traffic, and unifying security.

Catch up on everything from Connect 2026.

Watch keynotes and sessions on shipping agentic apps, governing AI traffic, and unifying security.

Check out the Cloudflare Connect Hub to follow along on the conference.

Connect 2026 sessions

All episodes

Thumbnail image for video "Poisoning the well: The rise of indirect prompt injection in the agentic age"

Poisoning the well: The rise of indirect prompt injection in the agentic age
As we shift from simple chatbots to autonomous AI agents that can browse the web, manage our calendars, and even audit our code, we have inadvertently opened a back door for a new breed of silent adversary. Indirect prompt injection (IDPI) allows...


How to get started on your FedRAMP journey with Cloudflare
Learn how to get started on your FedRAMP Moderate and FedRAMP High journey with Cloudflare in this interactive panel featuring Cloudflare’s Jake Schaeufele, Senior FedRAMP Compliance Specialist, and Tom Lianza, VP of Engineering. Going deep into the...


Automating app development the secure way: Velocity meets defense in depth
AI coding assistants have fundamentally transformed developer velocity. Today, developers can describe an app in plain English and deploy it to Cloudflare's global network in minutes. How do engineering teams keep that speed while ensuring every app...


Prevent cloud misconfigurations and data leakage before they happen
Cloud Workload Security is a new and unique approach to securing cloud resources. While most tools focus on catching risks after the fact, CWS uses an inline proxy to inspect cloud API calls as they happen. It enforces policy as code in real time,...


Bring your own logic with a fully programmable, secure web gateway
A fixed catalog of AI controls cannot anticipate every customer workflow, model, or risk signal. This session demonstrates how teams can extend Cloudflare Gateway policy with custom selectors and Cloudflare Workers invocations to build...


Better visibility and control: Shining a light on shadow AI
Blocking AI traffic on day one is a fast way to lose user trust and drive work into the shadows. This session shows how security teams can discover AI and MCP prompts automatically, run DLP in a non-blocking observation mode, measure what would be...


Auto model routing saves money, makes users happy, and is inherently flawed
Cloudflare AI Gateway includes an automatic model router that can reduce AI spending without reducing productivity. This talk walks through the router architecture that was built for performance, interpretability, and adaptability. But despite its...


The agentic network: Secure private connectivity for agents
AI agents can’t use VPN clients or type MFA codes, so every agentic workflow breaks at the same point: the agent needs to reach a private database, internal API, or MCP server. Exposing those resources publicly is a security risk, and legacy VPNs...


Behind the scenes: Challenges of running AI models at scale
Cloudflare runs at a massive scale — but what does scale mean when it comes to running AI models? We can talk about scale in different dimensions: global scale (routing, autoscaling, and concurrency), model catalog scale (breadth and modalities of...


Securing agentic harnesses: Control, visibility, and safe tool use
As organizations build agentic systems that call tools, retrieve data, and take action, security must move beyond prompt-level controls. This session explores how to secure agentic harnesses with the visibility, policy, and access controls needed to...
background