mnem-embed-providers

Embedding-provider adapters for mnem (OpenAI, Ollama). Sync, TLS-via-rustls, tokio-free.

Turns a user-configured provider into a concrete Embedder that the mnem CLI, MCP server, HTTP API, and Python bindings use to auto-embed node summaries on write and query strings on retrieve. All adapters are sync and built on [ureq] with rustls; mnem cannot afford to drag an async runtime into the CLI or the MCP server. API keys are never stored on disk - config records the name of the env var (api_key_env) and the key is read at adapter construction. The sibling-crate layout keeps mnem-core off any HTTP dependency, preserving the WASM-embeddability promise from ; see also for the trait-surface choice.

use mnem_embed_providers::{open, ProviderConfig, OpenAiConfig};

let cfg = ProviderConfig::Openai(OpenAiConfig {
    model: "text-embedding-3-small".into,
    ..Default::default
});
let embedder = open(&cfg)?;

Workspace top: ../../README.md.

Licensed under Apache-2.0.