v0.7.21

What's Changed

• Add timestamp insertion feature #176
• Add LINE Bot adapter with session management #177
• Fix timestamp_inserted_at handling in session state managers #178
• Refactor test assertions and update test messages #179
• Add timestamp options to server adapters #180
• Add LipSync and Blink to WebSocket example #181
• Update for v0.7.21 #182

Full Changelog: v0.7.19...v0.7.21

v0.7.19

What's Changed

• Support for OpenAI-compatible APIs #172
• Add support for custom model cid in Azure STT #173

Full Changelog: v0.7.18...v0.7.19

v0.7.18

What's Changed

• Add fallback for missing voice text tags in LLM response #167
• Add HTTP request retry logic to STT recognizers #168
• Add Request and Response Guardrail Support to LLM Services #169
• Update for v0.7.18 #170

Full Changelog: v0.7.17...v0.7.18

v0.7.17

What's Changed

• Update README.md #155
• Add user_id to LLM service methods #156
• Add speed parameter to SpeechGatewaySpeechSynthesizer #157
• Fix multi-character split delimiters #158
• Add Dify-compatible endpoint #161
• Update Claude model to claude-haiku-4-5 and fix tests #162
• Update for v0.7.17 #163

Full Changelog: v0.7.15...v0.7.17

v0.7.15

What's Changed

• Update default OpenAI model to gpt-5-search-api #149
• Add response chunk callback to HTTP server #150
• Support wave header detection in audio queue #151
• Add support for custom OpenAI module injection #152
• Add support for shared context histories in LLM services #153
• Update for v0.7.15 #154

Full Changelog: v0.7.13...v0.7.15

v0.7.13

What's Changed

• Add support for MCP tool integration #145
• Add built-in tools for web search and image generation #146
• Prevent HttpAdapter startup failure when resemblyzer is missing #147

Full Changelog: v0.7.12...v0.7.13

v0.7.12

What's Changed

• Add scenario name to evaluation #141
• Add experimental speaker registry with voice recognition #142
• Add speaker matching and naming to transcription API #143
• Update setup.py and README for v0.7.12 #144

Full Changelog: v0.7.11...v0.7.12

v0.7.11

What's Changed

• Add unified db_connection_str for database configuration #132
• Add extensible preprocess/postprocess to SpeechRecognizer #133
• Add experimental support for speaker diarization #134
• Mask unused tools in Dynamic Tool Call instead of removal in #135
• Fix bug where Gemini API fails without system instructions #136
• Add voice recorder directory support to adapters and pipeline #137
• Refactor adapter server constructors for STS pipeline config #138
• Add initial_messages support to LLM services #139
• Update setup.py for v0.7.11 #140

Full Changelog: v0.7.10...v0.7.11

v0.7.10

What's Changed

• Add on_recording_started callback to VAD detectors #122
• Add HTTP transcribe/synthesize endpoints #123
• Prevent errors with Azure OpenAI content filter streaming #124
• Add create_instant_synthesizer for HTTP-based TTS integration #125
• Add request merging to prevent fragmented speech recognition #126
• Add support for streaming PCM audio in chunks #127
• Remove extra newlines in merged request text #128
• Add volume_db_threshold to SileroSpeechDetector #129
• Add SessionStateManager for persistent session state management #130
• Update for v0.7.10 #131

Full Changelog: v0.7.9...v0.7.10

v0.7.9

What's Changed

• Add dialog evaluation API endpoints #117
• Add tool call tracking to dialog evaluation #118
• Add support for custom evaluation functions in DialogEvaluator #119

Full Changelog: v0.7.8...v0.7.9