SGLang, Azure OpenAI Service, Docling vs AnythingLLM

The registry records SGLang as available on self-hosted or private cloud or on-premise enterprise; it is scored here as self-hosted.

The registry records Azure OpenAI Service as available on vendor cloud or private cloud; it is scored here as enterprise SaaS.

The registry records Docling as available on self-hosted or private cloud; it is scored here as self-hosted.

The registry records AnythingLLM as available on self-hosted or private cloud; it is scored here as private cloud.

For this brief — confidential documents — self-hosted is a strong alternative. Change the brief and this verdict changes.

For this brief — confidential documents — enterprise SaaS is conditional: worth it only if the conditions in the options section hold. Change the brief and this verdict changes.

For this brief — confidential documents — self-hosted is a strong alternative. Change the brief and this verdict changes.

For this brief — confidential documents — private cloud is a strong alternative. Change the brief and this verdict changes.

Does not cover your main objective (document Q&A), only local LLM — so functional fit is capped at 0.25 however well the rest fits.

Covers your main objective (document Q&A) and 2 of 3 secondary objectives (internal company search, private company ChatGPT).

Covers your main objective (document Q&A) and 1 of 3 secondary objectives (internal company search).

Covers your main objective (document Q&A) and 3 of 3 secondary objectives (internal company search, local LLM, private company ChatGPT).

You stated no hosting preference, so no option is favoured on that basis. No hard restriction on where processing happens was stated. Headcount was not stated, so the size band could not be checked.

You stated no hosting preference, so no option is favoured on that basis. No hard restriction on where processing happens was stated. Headcount was not stated, so the size band could not be checked.

You stated no hosting preference, so no option is favoured on that basis. No hard restriction on where processing happens was stated. Headcount was not stated, so the size band could not be checked.

You stated no hosting preference, so no option is favoured on that basis. No hard restriction on where processing happens was stated. Headcount was not stated, so the size band could not be checked.

External data transfer: none. No catalogue deployment names SGLang, so this is read from “Private company knowledge base (self-hosted RAG)”, the best-matching self-hosted deployment for this brief. No edge in this design crosses out of the company network.

External data transfer: yes. No catalogue deployment names Azure OpenAI Service, so this is read from “Enterprise SaaS assistant with governance controls”, the best-matching enterprise SaaS deployment for this brief. Confidential content leaves your control on the Employees → Vendor assistant (web and desktop clients) link.

External data transfer: none. Computed from “Private company knowledge base (self-hosted RAG)”, the catalogue deployment that names Docling and best matches this brief. No edge in this design crosses out of the company network.

External data transfer: some. No catalogue deployment names AnythingLLM, so this is read from “Private-cloud RAG in a single region”, the best-matching private cloud deployment for this brief. Confidential content leaves your premises for your own cloud tenancy ("VPN / private endpoint"). You keep control of the account; the provider is a processor, so a DPA and a documented region apply.

Self-hosted: the content stays wherever you run the server, so residency is a property of your own infrastructure rather than a vendor commitment.

We hold no fetched residency commitment for Azure OpenAI Service. That is a question to put to the vendor, not an assumption to make about them. The vendor is headquartered in the United States, which is where a transfer question starts, not where it ends.

Self-hosted: the content stays wherever you run the server, so residency is a property of your own infrastructure rather than a vendor commitment.

We hold no fetched residency commitment for AnythingLLM. That is a question to put to the vendor, not an assumption to make about them.

Run on your own infrastructure, so the vendor questions apply only to whatever you still buy. We hold none of the four commitments for this subject, and for a self-hosted deployment none of them is required.

We have fetched none of the four commitments for Azure OpenAI Service: data processing agreement, subprocessor list, training on customer data, zero retention option. Each is a question for the vendor, and until it is answered it is unknown rather than absent.

Run on your own infrastructure, so the vendor questions apply only to whatever you still buy. We hold none of the four commitments for this subject, and for a self-hosted deployment none of them is required.

We have fetched none of the four commitments for AnythingLLM: data processing agreement, subprocessor list, training on customer data, zero retention option. Each is a question for the vendor, and until it is answered it is unknown rather than absent.

8–20 FDE-days to implement. No catalogue deployment names SGLang, so this is read from “Private company knowledge base (self-hosted RAG)”, the best-matching self-hosted deployment for this brief.

4–12 FDE-days to implement. No catalogue deployment names Azure OpenAI Service, so this is read from “Enterprise SaaS assistant with governance controls”, the best-matching enterprise SaaS deployment for this brief.

8–20 FDE-days to implement. Computed from “Private company knowledge base (self-hosted RAG)”, the catalogue deployment that names Docling and best matches this brief.

6–15 FDE-days to implement. No catalogue deployment names AnythingLLM, so this is read from “Private-cloud RAG in a single region”, the best-matching private cloud deployment for this brief.

You own GPU drivers and the model server, the identity integration and joiner/leaver process, operating-system patching and backups. Assessed difficulty 3/5 is the shape of that work; someone has to hold it after go-live.

The vendor runs the service. Your standing work is access control, the policy staff work under, and re-reading the contract and subprocessor list when they change — not patching or capacity.

You own GPU drivers and the model server, the identity integration and joiner/leaver process, operating-system patching and backups. Assessed difficulty 3/5 is the shape of that work; someone has to hold it after go-live.

You own GPU drivers and the model server, the identity integration and joiner/leaver process, operating-system patching and backups. Assessed difficulty 3/5 is the shape of that work; someone has to hold it after go-live.

3/5. Starting point 2/5: software you install, run and keep running on your own machines. +0.5 You run the model server yourself: GPU drivers, quantisation choice, memory headroom and restarts are all yours to own.

2/5. Starting point 1/5: a vendor product, configured rather than installed. +0.5 Single sign-on adds an identity provider, group-to-role mapping and a joiner/leaver process to the deployment.

3/5. Starting point 2/5: software you install, run and keep running on your own machines. +0.5 You run the model server yourself: GPU drivers, quantisation choice, memory headroom and restarts are all yours to own.

3/5. Starting point 2/5: your own tenancy to build in, but no hardware to buy or rack. +0.5 You run the model server yourself: GPU drivers, quantisation choice, memory headroom and restarts are all yours to own.

USD 10,000 – 48,000 — one-off — for SGLang at this headcount. No catalogue deployment names SGLang, so this is read from “Private company knowledge base (self-hosted RAG)”, the best-matching self-hosted deployment for this brief.

USD 3,100 – 23,000 — one-off — for Azure OpenAI Service at this headcount. No catalogue deployment names Azure OpenAI Service, so this is read from “Enterprise SaaS assistant with governance controls”, the best-matching enterprise SaaS deployment for this brief. Licences are not in that figure: no per-seat price has been read from the vendor’s own pricing page, so the software line is excluded from the total rather than guessed. Read it as implementation only.

USD 10,000 – 48,000 — one-off — for Docling at this headcount. Computed from “Private company knowledge base (self-hosted RAG)”, the catalogue deployment that names Docling and best matches this brief.

USD 5,200 – 30,000 — one-off plus monthly, read as first-year outlay — for AnythingLLM at this headcount. No catalogue deployment names AnythingLLM, so this is read from “Private-cloud RAG in a single region”, the best-matching private cloud deployment for this brief.

The registry files SGLang as open source, so the interface, retrieval behaviour and the model behind it can be changed — subject to the licence, which is a fetched fact on its own page.

Azure OpenAI Service is a vendor product: you configure what the vendor exposes — policies, connectors, retention settings — and nothing below that line.

The registry files Docling as open source, so the interface, retrieval behaviour and the model behind it can be changed — subject to the licence, which is a fetched fact on its own page.

The registry files AnythingLLM as open source, so the interface, retrieval behaviour and the model behind it can be changed — subject to the licence, which is a fetched fact on its own page.

7 issues spotted for this brief, 0 of them legal requirements; 0 carry a quote from the instrument they rest on. Leading with: Self-hosting moves the security obligation to you.

8 issues spotted for this brief, 0 of them legal requirements; 0 carry a quote from the instrument they rest on. Leading with: Confidentiality duties bind independently of data protection law.

7 issues spotted for this brief, 0 of them legal requirements; 0 carry a quote from the instrument they rest on. Leading with: Self-hosting moves the security obligation to you.

9 issues spotted for this brief, 0 of them legal requirements; 0 carry a quote from the instrument they rest on. Leading with: Confidentiality duties bind independently of data protection law.