diff --git a/README.md b/README.md
index b823821..19cce41 100644
--- a/README.md
+++ b/README.md
@@ -2,7 +2,8 @@
 
 This project provides an interface to communicate with the Thalmic Myo,
 providing the ability to scan for and connect to a nearby Myo, and giving access
-to data from the EMG sensors and the IMU.
+to data from the EMG sensors and the IMU. For Myo firmware v1.0 and up, access
+to the output of Thalmic's own gesture recognition is also available.
 
 The code is primarily developed on Linux and has been tested on Windows and
 MacOS.
@@ -16,6 +17,7 @@ understanding the protocol.
 
 - python >=2.6
 - pySerial
+- enum34 (for Python <3.4)
 - pygame, for the example visualization and classifier program
 - numpy, for the classifier program
 - sklearn, for a more efficient classifier (and easy access to smarter classifiers)
@@ -51,6 +53,14 @@ as the device name for the dongle; no argument means to auto-detect. You can
 also press 1, 2, or 3 on the keyboard to make the Myo perform a short, medium,
 or long vibration.
 
+To process the data yourself, you can call MyoRaw.add_emg_handler or
+MyoRaw.add_imu_handler; see the code for examples.
+
+If your Myo has firmware v1.0 and up, it also performs Thalmic's gesture
+classification onboard, and returns that information. Use MyoRaw.add_arm_handler
+and MyoRaw.add_pose_handler. Note that you will need to perform the sync gesture
+after starting the program (the Myo will vibrate as normal when it is synced).
+
 ## classify_myo.py (example pose classification and training program)
 
 classify_myo.py contains a very basic pose classifier that uses the EMG
@@ -74,7 +84,9 @@ had the chance to tinker with it yet.
 After you've done training with classify_myo.py, the Myo class in this file can
 be used to notify a program each time a pose starts. If run as a standalone
 script, it will simply print out the pose number each time a new pose is
-detected.
+detected. Use Myo.add_raw_pose_handler (rather than add_pose_handler) to be
+notified of poses from this class's classifier, rather than Thalmic's onboard
+processing.
 
 Tips for classification:
 
@@ -88,7 +100,7 @@ Tips for classification:
 # Caveats/issues
 
 - on Windows, the readings become more and more delayed as time goes on
-- doesn't have access to Thalmic's pose recognition
+- doesn't have access to Thalmic's pose recognition (for firmware < v1.0)
 - may or may not work with a Myo that has never been plugged in and set up with
   Myo Connect
 - classify_myo.py segfaults on exit under certain circumstances (probably