Long-term memory that personalizes the AI across conversations. Sparrow learns facts about you from chats and lets you add or remove them manually below — no setup required.
User Memories
Facts and preferences the AI has learned about you across conversations. Add or remove items below.
Chat Memory Embedding (advanced)
By default, chat memory uses the active embedding model from BYO Knowledge. Override here only if you want a different model for chat recall than for knowledge search.
Choose from your configured embedding models, or leave as default.
Select models to show in the chat model selector. Configure each provider below first, then add models here.
Add Model to Quick Access
Image generation models available on your plan. Models are managed by EchoNest — upgrade your plan for premium models.
Image Retention
How long to keep generated images stored before automatic deletion. Image URLs always refresh on access for the chosen retention period.
Existing images keep their original retention. New images use the value selected here.
Customize the AI assistant's behavior and personality by modifying the system prompt.
This is the instruction that shapes how the AI responds. The current date/time and available tools will be automatically appended.
This text will be automatically added to every message you send. Useful for persistent instructions like "use MCP tools" or "be concise".
Prompt Enhance (Magic Wand)
When enabled, the magic wand button will rewrite your prompt using prompt engineering best practices before sending.
Instructions that tell the AI how to rewrite your prompts. Customize this to match your workflow.
Choose which model rewrites your prompts. A fast, cheap model (e.g. GPT-4o-mini, Gemini Flash) works great here.
Document Generation
When enabled, the AI can produce downloadable PDF documents when you explicitly ask for a report, document or printable file. Turn off to remove the capability entirely; the model will never see the tool. Requires the Plus or Pro plan.
Only triggers when you explicitly ask for a document, report or PDF.
New look (beta)
Preview Sparrow's redesigned interface — a calmer, simpler layout with everything one tap away. Still in progress; you can switch back any time. All your features and data stay exactly the same.
Opt in to the upcoming UI redesign. Off by default.
Appearance
Choose how Sparrow looks. “System” follows your device’s light/dark setting.
One place for every model provider you connect to. Add as many named connections per provider as you need — different API keys, different servers, different regions — and pick which one is the default for each provider type.
Add connection
Used for embedding models and reranking. Get your key from Cohere Dashboard.
Define reusable instruction sets to shape AI behavior. Active skills are injected into every conversation — similar to .cursorrules or copilot-instructions.md.
Create Skill
Priority controls injection order (lower = first). Skill instructions are appended to the system prompt for every message.
Import / Export
Import skills from .md files or export your skills for backup and sharing.
Add MCP Server
Tool Settings
Maximum characters from MCP tool results. Larger = more context but slower responses. (1,000 - 100,000)
For power users who want full control over where embeddings + vectors live. When enabled, the chat-composer Knowledge button, chat memory and user memories route to your own embedding provider + Qdrant instead of our managed backend. Project file uploads and grounded Project chat are unaffected — they always use our managed S3 Vectors backend so cross-device search keeps working out of the box.
Routes the chat-composer Knowledge button, chat memory and user memories to your Qdrant using your active embedding model. Existing managed memories are preserved — toggle off to resume them. No data is migrated between backends.
Embedding Model
Add at least one embedding model to encode queries and documents. Only one model can be active at a time. Changing the active model requires re-indexing existing collections.
Qdrant Connection
URL of your vector database instance.
Retrieval Settings
Number of chunks to retrieve (1-200)
Minimum similarity score (0-1)
Maximum characters of context to include in prompt (500-100,000).
Metadata field name used to categorize collections in the dropdown. Leave empty to show flat list.
Rerank (Optional)
Reranking improves retrieval quality by reordering initial search results based on semantic relevance. Requires a Cohere API key configured in the Cohere provider tab.
When enabled, retrieved chunks are re-scored with a Cohere rerank model for sharper relevance.
Select the Cohere rerank model to use.
Number of top results to keep after reranking.