Key facts
| RAG endpoints | POST /v1/embeddings, POST /v1/rag/collections and POST /v1/rag/query |
| Embedding models | plugsky-embed-v1 (1536 dimensions, OpenAI-ada compatible) and plugsky-embed-large (3072 dimensions) |
| Retrieval modes | Keyword, vector and hybrid search with optional cross-encoder reranking |
| Citations | Every query returns ranked chunks with source attribution |
| Ingestion | Documents are chunked, embedded and indexed automatically per collection |
| Batch limits | Up to 2,048 inputs per embeddings request, max 8,191 tokens each |
| Data handling | Per-collection encryption at rest; API data is not used to train models |
| Deployment | Hosted, VPC, on-prem and air-gapped options; 30+ models behind one API |
TL;DR
- One retrieval pipeline, one collection and scoped key per client.
- Hold tenant keys server-side so rotation and revocation stay invisible.
- Flat monthly plans protect your margin; per-token billing makes it variable.
- Offer VPC, on-prem or air-gapped deployment for regulated accounts.
- Standardise onboarding and evaluation so each new client is repeatable.
How it works, step by step
- Define the service: which client documents, which questions, what turnaround.
- Design tenant isolation: collection, scoped key, rate limits and usage reporting.
- Build a reusable ingestion and evaluation checklist for onboarding.
- Keep tenant keys server-side in your gateway and attach tenant identity there.
- Configure per-collection encryption, audit logging and retention defaults.
- Offer private deployment options to clients with regulatory constraints.
- Review margin and model choice quarterly against real client usage.
Original data
Try it yourself
Open the RAG cost calculator →
Why MSPs need RAG grounded in their own content
MSPs win by productising what clients cannot staff. Document Q&A over client handbooks, runbooks, contracts and tickets is a natural managed service: every client has the corpus, few have the engineering time, and all of them care about isolation. The difference between a service and a liability is tenancy design.
Give each client a workspace concept: a collection for their documents, a scoped key your platform holds server-side, per-tenant rate limits and a usage report you can put on an invoice. Then the same retrieval pipeline serves every account without cross-tenant leakage.
The Plugsky RAG stack: embeddings, collections, queries
You do not have to assemble a vector database, an embedder, a retriever and a reranker separately. Documents are chunked, embedded and indexed automatically per collection, retrieval supports keyword, vector and hybrid modes, and every query returns ranked chunks with citations. The endpoints stay OpenAI-compatible — POST /v1/embeddings, POST /v1/rag/collections, POST /v1/rag/query — and the embeddings call also works standalone if you keep your own vector store. For MSPs, collections double as tenant boundaries: one per client, each with its own encryption scope and query surface, so the platform stays shared while the data does not.
The controls MSPs should configure first
Isolation is the product. Keep one collection per client, never place a tenant key in client-side code, and attach tenant identity in your own gateway so keys can be rotated invisibly. Per-collection encryption at rest, RBAC with workspace and resource-level permissions, and audit logs that record user, model and region give you answers when a client's security team asks.
For regulated accounts, offer VPC, on-prem or air-gapped deployment with BYOK, and review the DPA and sub-processor list as part of onboarding. Because self-serve plans are flat monthly with unlimited fair-use usage, your cost per client is predictable rather than metered per token.
A practical pilot path
Start with a single high-value document type — policies, onboarding handbooks or vendor contracts — and build the same evaluation set for every new client. Price the service, not the tokens. Prototype on the free plan, standardise the onboarding checklist, then move regulated clients to private deployment. See the live pricing page for current plans.
Honest comparison
| Capability | Plugsky | Assembling your own stack | Keyword search only |
|---|---|---|---|
| Retrieval | Keyword, vector and hybrid with optional reranking | You integrate and tune each component | Keyword matching, no semantics |
| Citations | Ranked chunks with source attribution | You build attribution yourself | Document links, not passages |
| Embeddings | plugsky-embed-v1 (1536d) and plugsky-embed-large (3072d) | Model choice spread across vendors | Not applicable |
| Deployment | Hosted, VPC, on-prem and air-gapped | Depends on every component | Usually on-prem already |
| Data handling | Per-collection encryption; API data not used to train models | You audit every vendor | Existing controls apply |
| Time to first answer | One collection and a query call | Weeks of integration work | Days, with lower answer quality |
Frequently asked questions
Can we white-label RAG to our clients?
Yes. Keep your own brand on the client-facing layer and use Plugsky as the retrieval and model platform behind it; the API is OpenAI-compatible and does not require client-facing branding.
How do we keep tenants separate?
Use a collection per client, scoped API keys held server-side, per-key rate limits and quotas, and audit logs that record which tenant produced each request.
Do we have to change our application code?
No. The RAG endpoints are OpenAI-compatible, so existing SDKs, prompts and evaluation harnesses keep working when you point the base URL at Plugsky.
Is there a free plan?
Yes — plugsky-micro and plugsky-lite with 2 API keys and no credit card. A 14-day full-access trial is also available.
How is pricing structured?
Self-serve plans are flat monthly with unlimited fair-use usage and no per-token charges or overage fees. See the live pricing page for current plans.
Does Plugsky train on our data?
Plugsky states that API data is not used to train models, and collections are encrypted at rest; on Enterprise you can keep everything inside your own VPC.