v0.9.9

What's Changed

• bump bolna version 0.9.7 -> 0.9.8 by @prateeksachan in #259
• Improve connection closures by @prateeksachan in #262
• push text responses for chat in tools by @prateeksachan in #263
• Azure speed rate by @prateeksachan in #273
• enable custom post content_type=form functions by @prateeksachan in #276
• latency: standardize per‑turn metrics across TTS + transcribers; aggregate in TaskManager by @san0808 in #274
• fix form urlencode for post requests by @prateeksachan in #278
• fix form urlencode for post requests by @prateeksachan in #279
• wait for transfer message by @prateeksachan in #282
• remove hardcoded schema by @pavankapoor31 in #288
• checking for multiple utterences by @prateeksachan in #285
• feat(transcriber): add AssemblyAI realtime transcriber by @san0808 in #280
• feat(transcriber): add Sarvam AI transcriber by @san0808 in #281
• temp: check handling for synthesizer schema by @prateeksachan in #290
• feat(transcriber): update SarvamTranscriber to support saaras and saarika models with separate endpoints by @san0808 in #289
• feat: Add Google Cloud Speech-to-Text transcriber by @san0808 in #286
• Feat/rag update v2 by @prateeksachan in #291
• fixing hard-coded raw log by @prateeksachan in #293
• feat: added naturalness to the rag agents during normal conv like sim… by @san0808 in #292
• refactor: Rename RAG service URL to RAG server URL across agents and … by @san0808 in #299
• adding logs by @prateeksachan in #300
• Feat/exotel integration by @san0808 in #301
• docs: fix API.md JSON examples and formatting by @sandipkumardey in #297
• docs: add minimal programmatic usage examples and runnable scripts by @sandipkumardey in #298
• tracking first audio and stream time by @prateeksachan in #304
• Dtmf integration by @san0808 in #307
• changes for cartesia sonic-3-preview by @prateeksachan in #309
• Latency tracking by @prateeksachan in #313
• typo by @prateeksachan in #315
• fix: fix and track each interim latency proeperly by @prateeksachan in #314
• using util function for timestamp by @prateeksachan in #317
• Fix latency format in knowledge-based and graph agents by @san0808 in #318
• handle exceptions by @prateeksachan in #319
• handling exception raised while normal closures by @prateeksachan in #320
• import fic by @prateeksachan in #321
• handle disconnection of chats gracefully by @prateeksachan in #322
• closing dg connections gracefully by @prateeksachan in #323
• fixing dg closures by @prateeksachan in #324
• Fix/turn latency by @prateeksachan in #325
• disconnection handling by @prateeksachan in #326
• Fix/turn latency by @prateeksachan in #329
• inject user and agent numbers by @prateeksachan in #330
• check for null in default vars by @prateeksachan in #331

New Contributors

• @san0808 made their first contribution in #274
• @pavankapoor31 made their first contribution in #288
• @sandipkumardey made their first contribution in #297

Full Changelog: 0.9.8...0.9.9

v0.9.8

What's Changed

• bump bolna version 0.9.6 -> 0.9.7 by @prateeksachan in #209
• removing keyterms for nova-3 by @prateeksachan in #210
• add error log for transcriber connect errors by @prateeksachan in #211
• update logic for hangup by @prateeksachan in #215
• incorporate voice_id for deepgram by @prateeksachan in #216
• update extraction prompt by @prateeksachan in #217
• adding rime TTS by @prateeksachan in #220
• incorporating sarvam websocket TTS by @prateeksachan in #222
• deepgram key fix by @prateeksachan in #229
• ifx logging for hangup model by @prateeksachan in #233
• poc forced initial message by @prateeksachan in #230
• remove redundant logs by @prateeksachan in #238
• adding elevenlabs host as env param by @prateeksachan in #239
• adding OpenRouter support by @prateeksachan in #241
• check if audio is still being played by @prateeksachan in #242
• passing custom params for live transfer_call by @prateeksachan in #247
• inject variables for user online message by @prateeksachan in #248
• removing log by @prateeksachan in #253
• Fix Deepgram websocket connection handling for authentication failures by @devin-ai-integration[bot] in #254
• Fix websocket ClientConnection closed attribute error by @devin-ai-integration[bot] in #255
• Disable LiteLLM INFO logs to improve performance by @devin-ai-integration[bot] in #256

New Contributors

• @devin-ai-integration[bot] made their first contribution in #254

Full Changelog: 0.9.7...0.9.8

v0.9.7

• adding smallest lightning-v2 model TTS https://github.com/bolna-ai/bolna/pull/206/files
• azure TTS latency optimization #191 h/t @abdulrahmanmajid
• tool calling support for litellm supported models #202
• docker updates, fixes and optimizations #194 h/t @belikewater2
• support for sarvam bulbul-2 TTS #179
• mocked responses for transfer call tool #177 in case of non-telephony conversations
• handle chat websocket closures gracefully #176

v0.9.6

• dict support and multiple param type support for tooling (#157)
• support for pre-tool custom message (#161)
• Upgrading websockets version to 15.0.1 (#167)
• Adding support for generate_precise_transcript for incorporating interruptions in transcripts (#170)
• Saving tool history for extraction (#171)
• Fix elevenlabs isFinal text chunker logic for non-alphanumeric edge cases (#172)
• Improve gradual disconnect on hangup letting the agent complete before hanging up (#173)
• Lots of bug fixes

v0.9.5

• web-call support over websockets (#155, #156)
• ElevenLabs TTS fix for isFinal (#142, #154)
• Adding language-specific filler message for function calls (#151)
• Updated OpenAI tooling capability (#137)
• Support for Azure STT (#138)
• Ingesting current time, timezone in agent context (#131)
• Lots of bug fixes

v0.9.4

• Bug fixes
• Using mark events for audio acknowledgements
• Improved support for elevenlabs latest models eleven_flash_v2_5
• Custom LLMs support

v0.9.3

• Standardization for audio formats in all Synthesizers
• Bug fixes
• Generate non-cached responses in semantic routes
• Supporting mulaw format for elevenlabs

v0.9.2

• Major bug fixes and improvements for async tasks
• OpenAI Function tooling
• Hangup using LLM-based prompts
• Improvements for Cartesia TTS and Smallest TTS
• Incorporate random uuid for run_ids
• Adding dynamic context to logs

v0.9.1

• Fixed typo for passing user provided synthesizer keys 90621fa

v0.9.0

• Enabling elevenlabs streaming audio
• Added support for Cartesia TTS
• Added support for Smallest TTS
• Audio format fixes and cleanup to remove unnecessary conversions
• bug fixes