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
- Collection boundaries should follow matters and confidentiality walls.
- Citations let lawyers verify every answer against the source passage.
- Scoped keys and audit logs make access attributable and reviewable.
- Review DPA, sub-processors and deletion workflows before privileged data moves.
- Pilot on template or precedent libraries before live matters.
How it works, step by step
- Map matters and confidentiality walls to collection boundaries.
- Define who can query which collection and issue scoped keys accordingly.
- Ingest a closed corpus first and validate chunking against contract structure.
- Require citations and build a question set with expected sources.
- Configure retention, deletion workflows and audit export with your IT and risk teams.
- Review the DPA, sub-processor list and SCCs before privileged content is loaded.
- Scale to live matters only after retrieval quality is demonstrated.
Original data
Try it yourself
Why legal teams need RAG grounded in their own content
Legal work is retrieval work. The value is in finding the right clause, the relevant precedent or the specific indemnity, and in being able to cite it. A model that answers without a source is unusable in practice; a model that surfaces the passage and points to the document is a research assistant.
That makes RAG the natural pattern for law firms and in-house teams. It also makes isolation the first design constraint: collection boundaries should follow matters or confidentiality walls, not convenience, and access should be scoped so that the wrong person cannot query the wrong corpus.
The Plugsky RAG stack: embeddings, collections, queries
The retrieval pipeline is three calls: embed, collect, query. POST /v1/embeddings converts text to vectors using plugsky-embed-v1 (1536 dimensions, OpenAI-ada compatible) or plugsky-embed-large (3072 dimensions); POST /v1/rag/collections stores and indexes documents with per-collection encryption at rest; POST /v1/rag/query returns ranked chunks with source attribution. Keyword, vector and hybrid retrieval are all available, with optional cross-encoder reranking when precision matters more than raw recall. For legal teams, matter collections and citations are what make retrieval defensible: the passage is the evidence, and the collection is the wall.
The controls legal teams should configure first
Design access before ingestion. Create a collection per matter or per confidentiality wall, issue scoped keys per team, and keep client data in a deployment that matches the engagement terms — hosted with a region pin, in your VPC, on-prem or air-gapped. Per-request audit logs record model, tokens, latency, user and region and export to your SIEM with retention up to 7 years, so access reviews have evidence.
Contract terms matter too: review the DPA, sub-processor list and EU SCC coverage before privileged material moves, and agree deletion workflows for closed matters. BYOK through your KMS or an on-prem HSM keeps encryption keys under the firm's control.
A practical pilot path
Begin with a closed evaluation: a template library, a settled matter or public precedent sets. Build a question set with expected sources, and score retrieval before drafting quality. Prototype on the free plan with plugsky-micro and plugsky-lite, then move to the deployment your engagement terms require. 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
How do we protect privilege?
Keep a collection per matter or client, scope keys to the team working on it, and deploy inside your VPC or on-prem where required. Audit logs then show who queried what and when.
What happens to data when a matter closes?
Agree deletion workflows up front: remove the collection and its documents, and confirm that logs and backups follow the retention you configured rather than an indefinite default.
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.