Supermemory vs Honcho

You send Supermemory content: messages, documents, URLs, PDFs, images, or video. It extracts facts and keeps a profile for each container tag, split into static facts like a user's role and dynamic ones like the project they're on now (user profiles). One profile() call returns it in about 50 milliseconds (README).

Supermemory links each new fact to older ones as an update, an extension, or a derived fact. An update "replaces what was true before for search purposes", history stays in the graph, and temporary facts drop after they expire (graph memory). Search is hybrid by default and returns memories and document chunks in one query.

Billing runs on four meters: text tokens, rich-content tokens, searches, and operations. Operations cover "counted platform operations not attributed to token or query meters", such as instant dreaming, reranking, and query rewriting (billing). Re-adding content under the same ID bills only what changed (pricing).

Honcho stores each message as it was sent, labelled with the peer who said it. A peer is any participant: a person or an agent. A background process called the deriver reasons over new messages and writes conclusions about each peer, including things that follow from what was said but were never stated outright. When a user corrects a fact, the deriver writes a new conclusion with a timestamp, and the old one stays for now. Dreaming runs later, once enough new conclusions have built up, deletes conclusions a newer fact has replaced, and merges ones that say the same thing.

Your agent reads memory two ways. session.context() returns the representation, a summary, and recent messages inside a token budget, in about 200 milliseconds, and those reads are unlimited. peer.chat() takes a question and reasons over the representation to answer it, at one of five levels from minimal to max, priced per query from $0.001 to $0.50.

Writes are asynchronous. A conclusion lands after the deriver runs, in time for the next turn but not the same one. Each peer can also keep its own view of every other peer it has observed. A support agent's model of a customer and the customer's model of that agent are separate objects.

Supermemory handles PDFs, images, video, and code, and syncs Google Drive, Gmail, Notion, OneDrive, and GitHub into the same API as its memory (README).

Supermemory's Scale plan includes SOC 2 and a HIPAA BAA (pricing). Honcho has SOC 2 Type 1 and doesn't offer HIPAA.

Supermemory's self-hosted server is one binary that needs only a model, with no Docker or database (self-hosting). Honcho's self-hosted stack runs the API, Postgres, and Redis on Docker.