memoryfeed

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    memoryfeed

    1

    Retrieve, search, and assemble cited context from your local MemoryFeed social and reading archive via MCP.

    Free

    1 useSecurity scanned

    memoryfeed

    Example session with this skill installed

    Search my memories from the last 30 days for any mentions of "vector databases" and summarize what I found with citations.

    • Read your context and instructions
    • Compiled the memoryfeed

    I have searched your MemoryFeed archive for the past 30 days. I found 4 relevant entries regarding vector databases, ranging from technical comparisons to implementation guides.

    Memory Retrieval: Vector Databases (Last 30 Days)

    DateSourceKey TakeawayCitation ID
    2023-10-15Twitter/XComparison of HNSW vs IVF indexing speeds for billion-scale vectors.[MF-882]
    2023-10-22NewsletterDiscussion on the trade-offs of pgvector for Postgres users.[MF-901]
    2023-10-28Technical BlogBenchmarking Qdrant and Weaviate for multi-tenant architectures.[MF-944]
    2023-11-02GitHub ReadmeArchitecture notes on local embedding storage and retrieval latency.[MF-962]

    Summary of Findings
    Your recent reading focuses heavily on performance trade-offs. You captured a comparison between HNSW and IVF indexing methods [MF-882] and looked into Postgres-native options via pgvector [MF-901]. More recently, your focus shifted to multi-tenant scaling [MF-944] and local latency optimization for embeddings [MF-962].

    Next steps

    1. Use build_memory_context to create a detailed briefing document using these specific citations.
    2. Run resurface_memory_context with these keywords to find related items older than 30 days.
    3. Check get_runtime_perf if you notice delays in indexing new vector-related bookmarks.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Local knowledge management often suffers from fragmented silos and poor recall. MemoryFeed bridges the gap between your local social content archive and your AI agent through a specialized MCP server. It allows agents to search, resurface, and assemble cited context from your personal reading history and captured social feeds.

    What it does

    • Targeted search Query your local memory archive with optional time-window filtering.
    • Context assembly Build bounded memory blocks with preserved numbered citations for downstream tasks.
    • Smart resurfacing Retrieve related older items based on relevance and heat-mapping.
    • Health monitoring Run diagnostic checks on your memory stack, project health, and runtime performance.
    • Timeline exploration Access chronological history across various platforms when server policies allow.

    How it works

    1. Initialize Verify the MemoryFeed MCP server is active and check stack health using detect_stack.
    2. Retrieve Use search_memory for specific queries or timeline_memories for date-based browsing.
    3. Analyze Process results via resurface_memory_context to identify high-interest historical data.
    4. Synthesize Generate cited reports or bounded context blocks using build_memory_context.

    Frameworks & tools

    Requires a local MemoryFeed installation running Python 3.12 or higher. The skill interfaces via the Model Context Protocol (MCP) using stdio transport. It supports optional integration with Gemini or Groq for enhanced processing.

    Why this beats prompting it yourself

    Writing raw prompts to interact with local databases often leads to context overflow or hallucinated citations. This skill enforces strict retrieval limits, maintains citation integrity, and manages the "heat" state of your memories to ensure relevance without manual tuning.

    Use cases

    • Find specific articles or social posts read during a specific week last month.
    • Assemble a cited summary of everything captured about a specific technical topic.
    • Audit the health and performance of your local memory indexing stack.
    • Surface forgotten but relevant bookmarks to inform a new writing project.

    Known limitations

    Requires an existing local MemoryFeed installation and MCP server configuration. Privacy boundaries prevent cloud transport unless explicitly enabled by the user.

    Requires local MemoryFeed installation and Python 3.12+. Remote transport requires manual configuration of TLS and authentication. Read-only by default.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 29 days ago

    • Free with an Agensi account

    Requires

    Python

    Permissions

    Read Files
    Write Files
    Terminal / Shell
    Listed29 days ago

    What's inside

    Frequently Asked Questions

    MemoryFeed solves the problem of fragmented local data by allowing AI agents to search, filter, and assemble cited context from your personal reading and social history via a structured MCP server. It ensures your agent provides accurate, cited information rather than hallucinating details from your past browsing.

    Yes, this skill requires an existing local MemoryFeed installation running on Python 3.12 or higher. It communicates via the Model Context Protocol (MCP) using stdio transport, meaning your agent must support MCP to interface with your archive.

    MemoryFeed is built for any AI agent or platform that supports the Model Context Protocol (MCP). While it works natively with standard MCP implementations, it offers enhanced processing capabilities when integrated with LLMs like Gemini or Groq.

    Security is a priority; all data remains within your local environment, and the skill utilizes private stdio transport for communication. No data is sent to the cloud unless you explicitly enable external transport policies in your server configuration.

    Unlike raw prompting, this skill manages 'memory heat' and relevance mapping to prevent context overflow. It also enforces strict citation integrity, ensuring every piece of retrieved information is tied back to a specific source in your archive.

    New in Agents & Orchestration