Hi, I’m Vivek Gupta, Founder & CEO at MemCode. Agent Device gives AI coding agents CLI, MCP, and a typed API for automating and verifying mobile apps across platforms. A durable memory layer could preserve app-specific interaction paths, device constraints, accepted test expectations, known recovery steps, and reviewer decisions across runs, while keeping live devices, credentials, screenshots, and test artifacts ephemeral or authoritative in their existing systems.
MemCode provides state-of-the-art semantic retrieval and durable updates with provenance, retention, export, and deletion controls. Agent Device could retrieve project- and app-scoped context before an agent performs a workflow and write back only explicit, approved findings or verified outcomes through MemCode’s MCP server or SDK. Local memory and self-hosted deployment are concrete options: the engine/API can run in the operator’s own Docker/VM or local process behind a private base URL, keeping records, embeddings, credentials, and logs inside that deployment with no mandatory hosted endpoint or telemetry. MCP + SDK support makes this an optional adapter without core-code changes.
I’d value the maintainers’ view on whether durable mobile-workflow context fits Agent Device’s roadmap. If useful, I can open a focused PR or discuss the smallest integration around the existing CLI/MCP boundary. The short overview is here: MemCode Leviathan.
Hi, I’m Vivek Gupta, Founder & CEO at MemCode. Agent Device gives AI coding agents CLI, MCP, and a typed API for automating and verifying mobile apps across platforms. A durable memory layer could preserve app-specific interaction paths, device constraints, accepted test expectations, known recovery steps, and reviewer decisions across runs, while keeping live devices, credentials, screenshots, and test artifacts ephemeral or authoritative in their existing systems.
MemCode provides state-of-the-art semantic retrieval and durable updates with provenance, retention, export, and deletion controls. Agent Device could retrieve project- and app-scoped context before an agent performs a workflow and write back only explicit, approved findings or verified outcomes through MemCode’s MCP server or SDK. Local memory and self-hosted deployment are concrete options: the engine/API can run in the operator’s own Docker/VM or local process behind a private base URL, keeping records, embeddings, credentials, and logs inside that deployment with no mandatory hosted endpoint or telemetry. MCP + SDK support makes this an optional adapter without core-code changes.
I’d value the maintainers’ view on whether durable mobile-workflow context fits Agent Device’s roadmap. If useful, I can open a focused PR or discuss the smallest integration around the existing CLI/MCP boundary. The short overview is here: MemCode Leviathan.