Android: Google Voice Recognition Server

I am currently playing around with Google's Voice Recognition API for Android SDK. What I want to know is the URL for their voice recognition. You need a data connection to use t

Solution 1:

Okay, here's what I found, building off of this article.

Here is a full TCP dump of the various HTTP and TCP operations that go into a voice search:

POST/m/voice-searchHTTP/1.1Content-Length:117Content-Type:application/octet-streamHost:www.google.comConnection:Keep-AliveUser-Agent:Mozilla/5.0(Linux;U;Android2.3.5;en-us;YP-G1)AppleWebKit/525.10+(KHTML,likeGecko)Version/3.0.4MobileSafari/523.12.2.....intent-speech-apiR\!VS2.1.3os=[Android2.3.5YP-G1]........>...".-16508905153536551372.en-US:......(.EZL..HTTP/1.1200OKContent-Type:application/x-protobufferDate:Tue,26Jun2012 13:07:17 GMTExpires:Tue,26Jun2012 13:07:17 GMTCache-Control:private,max-age=0X-Content-Type-Options:nosniffX-Frame-Options:SAMEORIGINX-XSS-Protection:1;mode=blockContent-Length:133Server:GSE76d72f95e7d9d47135f894ddb26d6303d....intent-speech-api..RH..74.125.142.126.....0ZQjyPfTdtfxMBro66d72f95e7d9d47135f894ddb26d6303dPOST/m/voice-search?multiproto=trueHTTP/1.1Content-Length:3271Content-Type:application/octet-streamHost:www.google.comConnection:Keep-AliveUser-Agent:Mozilla/5.0(Linux;U;Android2.3.5;en-us;YP-G1)AppleWebKit/525.10+(KHTML,likeGecko)Version/3.0.4MobileSafari/523.12.2.s.....intent-speech-apiRZ!VS2.1.3os=[Android2.3.5YP-G1]........>...".-16508905153536551372.en-US:......EZL...d.....intent-speech-apibK.j..free_form.............en-US(.0.:.com.rockwellcollins.voiceH.PVh.x...........intent-speech-api......<4p(.X.....*.EaU..74-l.`..X.M...<..a>.y..ioZ~Gr...F2JDNH.8..@...<(r'..~a..(B....4..,XxE..T.+...<.>cN.p.....O.j.V.3]d5q....J...<${c;.j.A..K.e..Gv.$@...C.....f@<..Z..B..i..^T]....*.c.Q.Ak.k...<,}_..b.....f"..CI.....D..f>2...<n].W......W.^).6........l...A@<.[m.........t.E.n.SC.D./`...<sx.tH.a...d#......U......[...@<.Co.........Ce.0,......}....:`<&p#......k...........?...!....<U.......3:.H.L....<j....:<..n.<"n.>....i....c]..,.2.<O{.W..,..........intent-speech-api......<.{%....a..:....)......../.Mq.@<.[.6.A.a..z9|p.......$...D""...<.r]..2.a...\.J...:.*.y..-.j0..P<$[c..\....O......j$..U}........<Dd_......Z.....x....^.!.....{..<$.cD.....r.rc...xh....4:.......<..}.]).a..[.5..K....d..=.=W..P<&..~.h...O..Qo..p~..'f!.n\.LN.<.._..........3..D5.~..Ws~.L..<*Cx..j.ai.Z.`R..\......h.c.....<.d....H...{zb..jo...L.o...$T.p<y.6.D@.Q.J.UY\.....]-.F.~.y....<C...5zQ.w.K...x3k...-.]....o{.<U/S..........QY.a.........4.\...........intent-speech-api......<,.c.......*.,......8....H.N."..<k.l.*{...k.....I$..?.B.....iQU.<%n$.m..t...2..Z......vT......{.<)...o..p.d<.../....`.H.v..=..~.<?%....R....Y...o..K.f.B.i|K...<Y.......T./h..z.g.@.......s\~.<a%..0R....?...........p..7..=Xp</...P^....Y}......z.........B.<>..........s2.....*.6..%.!Q...<.._....!.._..........Y.....}...<.edG.&..s..j....j.......-......<..X.AQ...=.4.9...k..r...Bx..f2P<F['.u........J............/...<.3S..r..........,.V.%...]....O..........intent-speech-api......<(}q..C/P.9.......zm.6s...)..]4P<..(..P......N.6...c......k..".<..x...q..>.....QI$[~...>..Q.L0.<..c......<..*..I.`........Jf..@<.v#F.......0R..BZ.I'..-...q..IP<..dN.tP!_Z...0.+a.j.29Yc..g...`<C.{D......Z2......7.l.2.F0....@<D._....a.*........+..5...N.yFP<D.d......}:.].-.%e......|-.p...<1..^.1.....=....E.U.z...g...@<..g..l..._/z...SH.A)]q....../..<&.i.\..a.{..{...../....w...B3..<..g........O*..<c...y...YE)...<$....p.A........)..OP....Y;.s.P.........intent-speech-api......<B}h....!.4.j(L.e.......F,..3#.<....NA...V........\V[.da....0<Deg......J.]Z..t...>@.$./...o.<..N.Z%..J.....x(wJ.l.5Y.=.oTSP<&ra..`....J.[....?.i$..d......p<s.......d.-..d...=.....L..j.`<D~f.....y.J......B..!...c..D.c.<..d?V......1.%.........u..}..K.<L.x..2...i:.=q6.].`.9.{..A..".p<.zWGVt.A..O!R....L.w.O.(8....Y.<DgcGT...n.c.s......}.C..*.rf< {t>.......8b.[..\."....<@..B?.<^$......,:Xl...%r..F....)O...<.#6.i)A...X.~Qc...TF.....[+.+..........intent-speech-api......<.nh.....h....h".......|wd..n.l.<(mk,.P.A....,...*.57..EG....p.0<....V............[ ..?]....lWQ@<0[q.VW.a.....n.....m.t..6..`...<.~y..r...d........}.....:...v..<.zb..c...-.2S@F.....oa.H.m..dD.<Rd_....A...%kr..d.g.8.m......<.egF.......[.S...N..W.........`<|#..........'..4.....(........<(_i&C:C.........MK..kd@}Jh..~..<.z...(.!y.5:m^..r..........AW.<.OkF..4..i:.._0.....A._........<E.y..{..*......M....=..!Q.A...<.Lg.^H.@...M7...d8.6.A3....q'.....~.....intent-speech-api..d`<0z]>.......`]...7Vp.%...K..V_@<.Mg>........D.t...N........8/..<(......A.9j..r....7.....-9.4............intent-speech-api......HTTP/1.1200OKContent-Type:application/x-protobufferDate:Tue, 26Jun2012 13:07:22 GMTExpires:Tue, 26Jun2012 13:07:22 GMTCache-Control:private, max-age=0X-Content-Type-Options:nosniffX-Frame-Options:SAMEORIGINX-XSS-Protection:1;mode=blockContent-Length:281Server:GSE.=70571f20e9244116a243b05ee852a7831....intent-speech-api..R........intent-speech-api..b.........testing.....D..?...test............testting............texting...........testemail.........."0571f20e9244116a243b05ee852a7831-1.,
&0571f20e9244116a243b05ee852a7831......Z.POST/m/voice-search?multiproto=trueHTTP/1.1Content-Length:257Content-Type:application/octet-streamHost:www.google.comConnection:Keep-AliveUser-Agent:Mozilla/5.0(Linux;U;Android2.3.5;en-us;YP-G1)AppleWebKit/525.10+(KHTML, likeGecko)Version/3.0.4MobileSafari/523.12.2.n.....voice-searchRZ!VS2.1.3os=[Android2.3.5YP-G1]........>...".-16508905153536551372.en-US:......EZL...B56d72f95e7d9d47135f894ddb26d6303d..intent-speech-api.......Z...K70571f20e9244116a243b05ee852a7831....intent-speech-api......0..R...Z..HTTP/1.1200OKContent-Type:application/x-protobufferDate:Tue, 26Jun2012 13:07:22 GMTExpires:Tue, 26Jun2012 13:07:22 GMTCache-Control:private, max-age=0X-Content-Type-Options:nosniffX-Frame-Options:SAMEORIGINX-XSS-Protection:1;mode=blockContent-Length:230Server:GSE.827a4a9198ae0fd36445660bf963609842....voice-search..R..<56d72f95e7d9d47135f894ddb26d6303d..intent-speech-api......>70571f20e9244116a243b05ee852a7831....intent-speech-api......,
&7a4a9198ae0fd36445660bf963609842......Z.

It's a lot to read, but basically what happens in the Android sends voice data over a TCP to 74.125.142.126 on port 19294 (in my case anyway). After that, a POST request is made to 74.125.225.176. Shortly thereafter, The Google Server (from the same IP) responds with a list of possible options, seen here as:

.testing.....D..?..
.test...........
.testting...........
.texting...........

The two URL's that seems to be involved are www.google.com/m/appreq/vs and www.google.com/m/voice-search.

That seems to be the gist of it. Basic HTTP action (GET/POST) requests. I'm not sure how the port 19294 fits into things, but I think it's the original transmission of voice data, which is processed and returned through HTTP. I hope this can help anyone who stumbles across the same problem.