solana-agent 17.0.4__tar.gz → 17.0.6__tar.gz

{solana_agent-17.0.4 → solana_agent-17.0.6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: solana-agent
-Version: 17.0.4
+Version: 17.0.6
 Summary: The Future of Work
 License: MIT
 Keywords: ai,openai,ai agents,agi
@@ -13,7 +13,7 @@ Classifier: Programming Language :: Python :: 3.12
 Classifier: Programming Language :: Python :: 3.13
 Classifier: Programming Language :: Python :: 3 :: Only
 Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
-Requires-Dist: openai (>=1.68.1,<2.0.0)
+Requires-Dist: openai (>=1.68.2,<2.0.0)
 Requires-Dist: pydantic (>=2.10.6,<3.0.0)
 Requires-Dist: pymongo (>=4.11.3,<5.0.0)
 Requires-Dist: zep-cloud (>=2.7.0,<3.0.0)
@@ -32,13 +32,30 @@ Description-Content-Type: text/markdown
 
 ![Solana Agent Logo](https://dl.walletbubbles.com/solana-agent-logo.png?width=200)
 
+## Agentic IQ
+
+Solana Agent is an AI Agent framework defined on five key dimensions:
+
+* **Multi-Modal Intelligence:** Agents process both text and audio seamlessly. Users interact naturally through their preferred communication channel.
+
+* **Episodic Memory:** Agents remember past conversations with users. This eliminates repetitive explanations and builds meaningful relationships.
+
+* **Cognitive Routing:** Queries go to the most qualified agent automatically. Each specialist has domain-specific knowledge for optimal responses.
+
+* **Value Alignment:** Every agent follows clear organizational values and ethics. This ensures consistent, principled responses aligned with human priorities.
+
+* **Extensible Tooling:** Agents recognize when specialized tools are needed. They extend capabilities beyond conversation to take actions via Python package integration.
+
+This approach creates AI Agents that are intelligent, contextually aware, and aligned with organizational values.
+
+
 ## Features
 
-* Multi-modal input-output streaming with text or audio by AI Agents
-* Conversational memory per user shared by all AI Agents
-* Routing based on AI Agent specializations
-* Organizational mission, values, goals, and guidance for all AI Agents
-* Robust AI Agent tool plugins based on standard python packages
+* Seamless text and audio streaming with real-time multi-modal processing
+* Persistent memory that preserves context across all agent interactions
+* Intelligent query routing to agents with optimal domain expertise
+* Unified value system ensuring aligned, principled agent responses
+* Powerful tool integration that extends capabilities using standard Python packages
 
 ## Stack
 
@@ -109,17 +126,20 @@ async for response in solana_agent.process("user123", "What are the latest AI de
 ```
 
 ## Models Used
-* The model used for AI Agents is `gpt-4o-mini`
-* The model used for internal structured outputs is `gpt-4o-mini`
-* The model used for audio_transcription is `gpt-4o-mini-transcribe`
-* The model used for tts is `tts-1`
+* `gpt-4o-mini`
+* `gpt-4o-mini-transcribe`
+* `tts-1`
 
 ## Solana Agent Kit
 
+A collection of Solana Agent tools
+
 [Solana Agent Kit](https://github.com/truemagic-coder/solana-agent-kit)
 
 ## Example App
 
+A Solana Agent example app written in FastAPI and Next.js
+
 [Solana Agent Example App](https://github.com/truemagic-coder/solana-agent-app)
 
 ## License

{solana_agent-17.0.4 → solana_agent-17.0.6}/README.md

@@ -9,13 +9,30 @@
 
 ![Solana Agent Logo](https://dl.walletbubbles.com/solana-agent-logo.png?width=200)
 
+## Agentic IQ
+
+Solana Agent is an AI Agent framework defined on five key dimensions:
+
+* **Multi-Modal Intelligence:** Agents process both text and audio seamlessly. Users interact naturally through their preferred communication channel.
+
+* **Episodic Memory:** Agents remember past conversations with users. This eliminates repetitive explanations and builds meaningful relationships.
+
+* **Cognitive Routing:** Queries go to the most qualified agent automatically. Each specialist has domain-specific knowledge for optimal responses.
+
+* **Value Alignment:** Every agent follows clear organizational values and ethics. This ensures consistent, principled responses aligned with human priorities.
+
+* **Extensible Tooling:** Agents recognize when specialized tools are needed. They extend capabilities beyond conversation to take actions via Python package integration.
+
+This approach creates AI Agents that are intelligent, contextually aware, and aligned with organizational values.
+
+
 ## Features
 
-* Multi-modal input-output streaming with text or audio by AI Agents
-* Conversational memory per user shared by all AI Agents
-* Routing based on AI Agent specializations
-* Organizational mission, values, goals, and guidance for all AI Agents
-* Robust AI Agent tool plugins based on standard python packages
+* Seamless text and audio streaming with real-time multi-modal processing
+* Persistent memory that preserves context across all agent interactions
+* Intelligent query routing to agents with optimal domain expertise
+* Unified value system ensuring aligned, principled agent responses
+* Powerful tool integration that extends capabilities using standard Python packages
 
 ## Stack
 
@@ -86,17 +103,20 @@ async for response in solana_agent.process("user123", "What are the latest AI de
 ```
 
 ## Models Used
-* The model used for AI Agents is `gpt-4o-mini`
-* The model used for internal structured outputs is `gpt-4o-mini`
-* The model used for audio_transcription is `gpt-4o-mini-transcribe`
-* The model used for tts is `tts-1`
+* `gpt-4o-mini`
+* `gpt-4o-mini-transcribe`
+* `tts-1`
 
 ## Solana Agent Kit
 
+A collection of Solana Agent tools
+
 [Solana Agent Kit](https://github.com/truemagic-coder/solana-agent-kit)
 
 ## Example App
 
+A Solana Agent example app written in FastAPI and Next.js
+
 [Solana Agent Example App](https://github.com/truemagic-coder/solana-agent-app)
 
 ## License

{solana_agent-17.0.4 → solana_agent-17.0.6}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "solana-agent"
-version = "17.0.4"
+version = "17.0.6"
 description = "The Future of Work"
 authors = ["Bevan Hunt <bevan@bevanhunt.com>"]
 license = "MIT"
@@ -18,7 +18,7 @@ python_paths = [".", "tests"]
 
 [tool.poetry.dependencies]
 python = ">=3.12,<4.0"
-openai = "^1.68.1"
+openai = "^1.68.2"
 pydantic = "^2.10.6"
 pymongo = "^4.11.3"
 zep-cloud = "^2.7.0"

{solana_agent-17.0.4 → solana_agent-17.0.6}/solana_agent/services/query.py

@@ -94,32 +94,12 @@ class QueryService(QueryServiceInterface):
             agent_name = await self.routing_service.route_query(user_text)
             print(f"Routed to agent: {agent_name}")
 
-            # For audio mode, we need to carefully handle the response to make sure
-            # tool calls are properly executed and formatted for audio
+            # Generate response
             if output_format == "audio":
-                # Use the agent service to generate the response
-                text_response = ""
-
-                # First, get complete text response
-                # Note: This is a separate call from the audio generation
-                temp_response = ""
-                async for chunk in self.agent_service.generate_response(
-                    agent_name=agent_name,
-                    user_id=user_id,
-                    query=user_text,
-                    memory_context=memory_context,
-                    output_format="text"
-                ):
-                    temp_response += chunk
-
-                # Store the complete text for memory
-                text_response = temp_response
-
-                # Now generate audio from same request
                 async for audio_chunk in self.agent_service.generate_response(
                     agent_name=agent_name,
                     user_id=user_id,
-                    query=user_text,
+                    query=query,
                     memory_context=memory_context,
                     output_format="audio",
                     audio_voice=audio_voice,
@@ -129,15 +109,13 @@ class QueryService(QueryServiceInterface):
                 ):
                     yield audio_chunk
 
-                # Store conversation in memory
-                if self.memory_provider and text_response:
+                if self.memory_provider:
                     await self._store_conversation(
                         user_id=user_id,
                         user_message=user_text,
-                        assistant_message=text_response
+                        assistant_message=self.agent_service.last_text_response,
                     )
             else:
-                # For text mode, we can collect the response and store it directly
                 full_text_response = ""
                 async for chunk in self.agent_service.generate_response(
                     agent_name=agent_name,
@@ -149,7 +127,6 @@ class QueryService(QueryServiceInterface):
                     yield chunk
                     full_text_response += chunk
 
-                # Store conversation in memory
                 if self.memory_provider and full_text_response:
                     await self._store_conversation(
                         user_id=user_id,