7d730baa4c
* chore(config): remove deprecated ai provider and add memory readme section * docs(handoff): set PR number 106 * docs(project-map): update after structural changes --------- Co-authored-by: opencode-agent <agent@opencode.local>
40 lines
No EOL
1.3 KiB
Text
40 lines
No EOL
1.3 KiB
Text
# OpenAI Embeddings (Memory CLI)
|
|
# OpenRouter defaults (Qwen3 8B, batch=50). For OpenAI direct, use:
|
|
# OPENAI_BASE_URL=https://api.openai.com/v1
|
|
# OPENAI_EMBEDDING_MODEL=text-embedding-3-small
|
|
# OPENAI_EMBEDDING_BATCH_SIZE=2048
|
|
OPENAI_BASE_URL=https://openrouter.ai/api/v1
|
|
OPENAI_API_KEY=your-openrouter-api-key
|
|
OPENAI_EMBEDDING_MODEL=qwen/qwen3-embedding-8b
|
|
OPENAI_EMBEDDING_BATCH_SIZE=50
|
|
OPENAI_EMBEDDING_BATCH_DELAY=0.5
|
|
MEMORY_CHUNK_SIZE=512
|
|
MEMORY_CHUNK_OVERLAP=64
|
|
|
|
# OpenCode Server
|
|
OPENCODE_SERVER_PASSWORD=your-opencode-server-password
|
|
|
|
# GitHub
|
|
GITHUB_TOKEN=your-github-token-here
|
|
|
|
# Context7 MCP
|
|
CONTEXT7_API_KEY=your-context7-api-key-here
|
|
|
|
# Telegram (optional)
|
|
TELEGRAM_API_ID=your-telegram-api-id
|
|
TELEGRAM_API_HASH=your-telegram-api-hash
|
|
TELEGRAM_BOT_TOKEN=your-telegram-bot-token
|
|
|
|
# Cloudflare Tunnel (optional — see separate tunnel repo)
|
|
CLOUDFLARE_TUNNEL_TOKEN=
|
|
TUNNEL_DOMAIN=
|
|
|
|
# Antidetect Browser MCP
|
|
# Local: http://localhost:8765/mcp
|
|
# Remote: http://<your-server-ip>:8765/mcp
|
|
ANTIDETECT_BROWSER_MCP_URL=http://localhost:8765/mcp
|
|
|
|
# OpenCode Memory (optional — auto-setup on first memory-save if unset)
|
|
# Point at your fork of opencode-memory to sync via git push
|
|
OPENCODE_MEMORY_REMOTE=https://github.com/your-username/your-opencode-memory.git
|
|
OPENCODE_MEMORY_DIR=/root/.local/share/opencode/opencode-memory |