AnythingLLM 1.16.0
A single-container knowledge base and chat application, deployed and measured against the same corpus as every other platform here.
In one paragraph
Deployed from its published image, AnythingLLM runs 1 container and held 542.2 MB at peak during ingestion. It retained 98% of the facts planted in the corpus and answered 83% of the retrieval questions correctly.
Measured
| Measurement | Result | How |
|---|---|---|
| Time to first served request | 146 s | Includes image pull, from a clean host |
| Containers | 1 | Running after the stack settles |
| Idle memory | 258.1 MB | Sum across containers, 60s after ready |
| Peak memory during ingestion | 542.2 MB | Sampled every 5s across the whole corpus |
| Documents ingested | 36/36 | Failures counted, not excluded |
| Chunks stored | 1,900 | Across the whole corpus |
| Parsing fidelity | 98% | Approximate: no chunk-listing API, measured by exhaustive retrieval |
| Retrieval accuracy | 83% | Answer present in top-5 context, 93 questions |
| Retrieval latency | 274.3 / 307.2 ms | p50 / p95, CPU embedding |
| Licence | MIT | Permits offering it as a service |
What it does well
- One container and a few hundred megabytes at idle — an order of magnitude lighter than the purpose-built platforms
- The fastest path from nothing to a working knowledge base of anything measured
- MIT licensed, with no restriction on commercial use or resale
Where it falls short
- Ships wide open — the default deployment runs with authentication disabled, and the endpoint that mints a full-access API key answers without any credential at all
- Model settings are written through one untyped environment endpoint, and a numeric value where it expects a string crashes the handler with a bare HTTP 500
- Exposes no API listing a document's stored chunks, so parsing fidelity here can only be approximated through retrieval
- Uploading a document does not attach it to a workspace; that is a separate call, and without it the document is never embedded
Choose it when
- Small teams or single users who want a knowledge base running in minutes
- Hosts where memory is the binding constraint
- Anyone who needs a permissive licence
Choose something else when
- Anything exposed to a network without putting authentication in front of it first
- Buyers who need auditable visibility into how documents were chunked
Deployment notes
The lightest deployment measured by a wide margin: one container, a few hundred megabytes, ready in a couple of minutes. If the comparison were only about cost to run, this would win it outright.
The default security posture
Worth stating plainly because it is easy to miss: a fresh deployment reports
RequiresAuth: false and MultiUserMode: false, and
POST /api/system/generate-api-key returns a working key to an unauthenticated
caller. That is convenient for automation and dangerous for anything reachable
from a network. Put authentication in front of it before exposing the port.
Why its parsing figure carries a tilde
There is no route that lists the chunks stored for a document, so the parsing measurement is taken by sweeping retrieval with a wide top-N and collecting what comes back. That can only under-report: a chunk retrieval never surfaces is invisible to the measurement. A missing fact here means not observed, not dropped.