The goal of voiceglue is to be VXML 2.0 compliant, with as many VXML 2.1 features as possible.

Things known to be unsupported at this version:

• Speech recognition
• Recording of audio formats other than raw ulaw (”audio/basic”)
• SSML speech markup
• Changing timeout properties of audio file fetch (fetchtimeout, maxage, etc.)
• semantic interpretation of DTMF grammars
• Complex SRGS DTMF grammars (those with more than 1 variable length section)

Several of these are close to being supported, I just haven’t had time to finish them off yet.

No doubt there are other (hopefully small) areas of VXML that aren’t fully supported, and I would appreciate knowing about them on the voiceglue mailing list.