MemData is a decentralized memory infrastructure designed for autonomous AI agents and developers seeking persistent, searchable storage of unstructured data. The platform enables agents and applications to ingest files—including PDFs, images, screenshots, audio, and text—automatically handling optical character recognition (OCR) for images and PDFs, as well as audio transcription. Uploaded content is chunked, embedded, and indexed without the need for manual configuration or tuning, allowing users to query their data in natural language and receive relevant context with source citations.
MemData is accessible via a REST API and supports integration with AI tools such as Claude, ChatGPT, Gemini, Cursor, and automation platforms like n8n and Make. Developers can interact with the service using API calls or the MCP server, and file ingestion and querying are demonstrated through simple command-line examples. The platform supports a variety of file types, including PDFs, PNG, JPG, MP3, WAV, M4A, and text files, and provides persistent, account-isolated storage.
3 and at rest with AES-256. Data is stored on US-based infrastructure using SOC 2 compliant providers, and users retain ownership of their data, which is not used for model training or sold. Data can be deleted at any time through the API or dashboard, with purging completed within 24 hours.
MemData offers a tiered pricing model. The Free plan includes 100 MB of storage, 250 queries per month, a 10 MB file size limit, and access to OCR, audio transcription, and both API and MCP interfaces. The Pro plan, at $29 per month, expands storage to 10 GB, allows 10,000 queries per month, and increases the file size limit to 100 MB, adding priority support. The Scale plan, at $99 per month, provides 100 GB of storage, 50,000 queries per month, a 500 MB file size limit, priority support, and custom integrations. No credit card is required to start, and users can upgrade as needed.
Positioned as a complete memory pipeline rather than just a vector database, MemData is built for AI builders and agents requiring long-term, semantic memory and context retrieval capabilities.
In the RAG, search & retrieval space, MemData takes a focused approach. It focuses on providing persistent, searchable memory and context retrieval for autonomous AI agents and applications. It is built as a B2B product for AI agent developers. It ships for the web, embeddable surfaces, API, and the command line.
MemData builds and maintains MemData, and it first shipped in 2026. Among its 8 catalogued features are semantic memory, API access, and file ingestion. It exposes integrations via a public API and an MCP server.
Summary written by a language model from the project’s public pages.
What PulseGate has recorded for this listing
Closest matches by what these projects do