Key facts
| API compatibility | OpenAI-compatible chat, embeddings and function calling |
| Use cases | NOC runbook Q&A, field engineer assistance, support triage, incident summaries |
| Data controls | Scoped API keys, RBAC, SSO, audit logs and region selection |
| Performance | Streaming and function calling are live for interactive operations tooling |
| Deployment | Plugsky cloud, your VPC, on-prem and air-gapped options |
| Models | 30+ models from free aliases to frontier reasoning |
| Pricing model | Flat monthly self-serve plans with unlimited fair-use usage |
| Free tier | plugsky-micro and plugsky-lite on the free plan, no card required |
TL;DR
- Put runbooks and past incidents behind retrieval with citations.
- Give field engineers an assistant that answers in plain language.
- Triage subscriber tickets by intent and route them to the right queue.
- Summarise incidents into a consistent timeline for postmortems.
- Keep subscriber data out of prompts and scope keys per system.
How it works, step by step
- Choose one operational workflow, such as NOC runbook Q&A or field procedure lookup.
- Index runbooks, procedures and sanitised incident records with stable document ids.
- Require citations so an engineer can verify the source before acting.
- Triage support tickets by intent and route them, keeping human review for account changes.
- Scope keys per system (NOC, CRM, field app) and enable audit logging before rollout.
- Measure answer accuracy and mean time to resolution, then expand to incident summarisation.
Try it yourself
Telecom workflows that benefit first
Telecoms run on procedures and incident history, which makes grounded retrieval the right starting point:
- NOC knowledge: answer runbook questions with citations so on-call engineers can verify quickly.
- Field assistance: surface the right procedure for an asset, alarm or site condition.
- Support triage: classify tickets by intent, service and urgency, then route to the right queue.
- Incident summarisation: turn logs and notes into a consistent timeline for postmortems.
Never let an assistant execute changes; it recommends, an engineer acts.
Security and subscriber data
Subscriber identity, usage records and location data are highly sensitive, and many jurisdictions impose specific handling rules. Keep those fields out of prompts unless a documented, lawful process requires them, and prefer sending internal references and sanitised text. Configure scoped API keys per system, RBAC and SSO for staff, audit logs for attribution, and region selection for processing. Where regulation or a partner agreement requires stricter control, move the workload to a VPC, on-prem or air-gapped deployment. Plugsky supplies the controls; your legal and security teams define the boundary.
Latency, scale and operations
Interactive tooling needs predictable response times, while reporting and summarisation can run asynchronously. Split the two: streaming chat for engineer-facing assistants, batch-style jobs for incident and ticket summarisation. Route classification to small models and reserve frontier models for complex diagnosis. Because Plugsky is OpenAI-compatible, the same integration code serves cloud, VPC and on-prem deployments, and you can A/B test models by changing a model name. Keep an evaluation set of real (sanitised) questions so model swaps are measured, not assumed.
Pilot to production
Start on the free plan with plugsky-micro or plugsky-lite on a single NOC team and a runbook corpus. Use the 14-day full-access trial to compare models on hard cases, scoring citation support, accuracy and time-to-answer. Self-serve plans are flat monthly with fair-use usage, so growing ticket volume does not create billing surprises. Roll out per team with versioned prompts, keep an escalation path to senior engineers, and re-run your evaluation set whenever runbooks or models change.
Honest comparison
| Capability | Plugsky | Typical per-token API | Building in-house |
|---|---|---|---|
| API compatibility | OpenAI-compatible chat, embeddings and tools | Usually compatible | Full rewrite |
| Deployment | Cloud, VPC, on-prem and air-gapped | Mostly cloud-only | You operate GPUs and serving |
| Data residency | Region selection for each market | Often limited regions | You control fully |
| Pricing | Flat monthly self-serve, fair-use usage | Per-token, harder to forecast | GPU plus operations cost |
| Model choice | 30+ models behind one API | Varies by provider | You host every model |
Frequently asked questions
Can we keep our existing OpenAI SDK code?
Yes. Plugsky exposes an OpenAI-compatible API, so migration is a base URL and model-name change.
Is there a free plan?
Yes. The free plan includes plugsky-micro and plugsky-lite with no credit card, enough for a single-team NOC pilot.
Can the assistant change network configuration?
No. Keep the assistant to retrieval, drafting and recommendations; changes should go through your existing approval and change-control process.
How do we handle subscriber data?
Keep subscriber identity, usage and location fields out of prompts unless a documented lawful process requires them, and prefer internal references over raw records.
Can we deploy on-prem?
Yes. VPC, on-prem and air-gapped deployments are available, using the same compatible API and application code.
How does pricing work?
Self-serve plans are flat monthly with unlimited fair-use usage. See the live pricing page for current plans and enterprise options.
Which endpoints are live today?
Chat, streaming, JSON mode, function calling, embeddings, RAG and agents are live. Audio, images, moderation, files, batch, fine-tuning, assistants and responses are coming soon.