Repository navigation

Integrate Cognee memory engine with SiliconFlow LLM provider - #61

Closed

HeDaas-Code with Copilot wants to merge 25 commits into

copilot/refactor-deepagents-optimizationfrom

copilot/optimize-cognee-memory-system

Closed

Integrate Cognee memory engine with SiliconFlow LLM provider#61
HeDaas-Code with Copilot wants to merge 25 commits into
copilot/refactor-deepagents-optimizationfrom
copilot/optimize-cognee-memory-system

Commits on Feb 19, 2026
- Initial plan
  Copilot
  committed
  e98650a
- Add Cognee dependency and create cognee_memory.py, worldview_builder.py
  
  Copilot
  and
  HeDaas-Code
  committed
  3476998
- Integrate Cognee memory system and worldview builder with GUI
  
  Copilot
  and
  HeDaas-Code
  committed
  b40b405
- Add Cognee integration tests and fix database test configuration
  
  Copilot
  and
  HeDaas-Code
  committed
  3b5e916
- Update README English section with Cognee and Worldview features
  
  Copilot
  and
  HeDaas-Code
  committed
  9b7fc46
Commits on Feb 20, 2026
- Fix code review issues: API key priority, event loop leaks, unused imports, NPC keyword
  
  Copilot
  and
  HeDaas-Code
  committed
  030af1a
- Improve JSON parsing robustness and add raw output logging for debugging
  
  Copilot
  and
  HeDaas-Code
  committed
  661b9e8
- Configure Cognee to use SiliconFlow as custom LLM provider instead of OpenAI
  
  Copilot
  and
  HeDaas-Code
  committed
  f07e34f
- Improve error logging in CogneeMemoryManager to show detailed exception info
  
  Copilot
  and
  HeDaas-Code
  committed
  ca16b48
- Improve Cognee configuration: add timeout, better logging, API key validation
  
  Copilot
  and
  HeDaas-Code
  committed
  ca7f2e0
- Fix Cognee LLM configuration: use set_llm_config for complete setup, increase timeout to 120s
  
  Copilot
  and
  HeDaas-Code
  committed
  37425f5
- Fix Cognee config: use LLMConfig and EmbeddingConfig classes with SiliconFlow custom provider
  
  Copilot
  and
  HeDaas-Code
  committed
  a659a60
- Fix Cognee model names: use SiliconFlow format directly without openai/ prefix
  
  Copilot
  and
  HeDaas-Code
  committed
  d4463c9
- Fix Cognee config: use environment variables only (compatible with Cognee 0.5.x)
  
  Copilot
  and
  HeDaas-Code
  committed
  6494942
Commits on Feb 21, 2026
- Fix LiteLLM provider prefix: add openai/ prefix for SiliconFlow models
  
  Copilot
  and
  HeDaas-Code
  committed
  104428a
- Make LiteLLM model prefix configurable via COGNEE_MODEL_PREFIX env var
  
  Copilot
  and
  HeDaas-Code
  committed
  d1b8afd
- Remove misleading openai/ prefix, use model names directly for custom providers
  
  Copilot
  and
  HeDaas-Code
  committed
  ab2db3a
- Add LLM_INSTRUCTOR_MODE config for Cognee structured output control
  
  Copilot
  and
  HeDaas-Code
  committed
  3d3c8b3
- Fix Cognee LLM config: use openai provider with OPENAI_API_BASE for SiliconFlow
  
  Copilot
  and
  HeDaas-Code
  committed
  65d3aa5
- Fix embedding tokenizer error: add HUGGINGFACE_TOKENIZER config for SiliconFlow models
  
  Copilot
  and
  HeDaas-Code
  committed
  f851640
- Fix embedding tokenizer: register model variants in litellm.embedding_model_tokenizer_mapping
  
  Copilot
  and
  HeDaas-Code
  committed
  e33e97e
- Add log_warning method to DebugLogger class
  
  Copilot
  and
  HeDaas-Code
  committed
  c40c52d
- Fix tiktoken tokenizer error: register model variants in tiktoken.model.MODEL_TO_ENCODING
  
  Copilot
  and
  HeDaas-Code
  committed
  ab1ce79
- Add HuggingFace tokenizer with SSL fix and mirror support for China
  
  Copilot
  and
  HeDaas-Code
  committed
  8ee9ed7
- Add standalone Cognee manager app with log viewer, remove tiktoken
  
  Copilot
  and
  HeDaas-Code
  committed
  184a165

Commits

Commits on Feb 19, 2026

Commits on Feb 20, 2026

Commits on Feb 21, 2026