Real-time video rate adaptation has become an active topic with the arrival of video communication through the web browser. The most used rate adaptation algorithm in this area is one developed by Google and is being used in Chromium (and Google Chrome). That algorithm, as well as other alternative frameworks and algorithms available today, have a tendency to focus on a specific metric. At high level, the set of requirements for a rate adaptation system includes quick reaction times, fairness to other network traffic and good performance in heterogeneous networks. This thesis presents a comparison of some current adaptation systems and a list of requirements and metrics for such systems. Also, the design of a new framework for rate adaptation and congestion detection is presented, where the main focus is to use several different metrics together to detect network congestion and compute an appropriate sending rate. A test implementation with focus on delay based adaptation was developed and used in a series of tests to compare it with Google’s algorithm and to draw conclusions of performance and design decisions. The results indicate that the test implementation was able to compete with Google’s algorithm in the given test cases. It also revealed the importance of making good adaptation decisions based on the detected issues. Weaknesses of existing strategies are revealed and it is discussed how these can be solved.