Key facts
| API compatibility | OpenAI-compatible /v1/chat/completions; drop-in base URL change |
| Models | 30+ models behind one API; open-weight options for private deployment |
| Typical patterns | Disruption messaging, policy Q&A, operations knowledge search, cargo documents |
| Integration path | Connects to CRM, reservations and contact-centre platforms via middleware |
| Pricing model | Flat monthly self-serve plans with unlimited fair-use usage; no per-token billing — see the live pricing page |
| Free tier | plugsky-micro and plugsky-lite on the free plan, no card required |
| Deployment | Plugsky cloud, your VPC, on-prem and air-gapped options |
| Identity and audit | SSO/SCIM, role-based access, SIEM audit log export and BYOK via KMS or HSM |
TL;DR
- Ground every passenger message in current policy, with agent approval.
- Keep PNR and payment fields to the minimum the workflow needs.
- Choose region selection and audit export for cross-border obligations.
- Keep safety-critical decisions outside AI scope; plan a degraded mode.
- Start free with plugsky-micro and plugsky-lite, no card required.
How it works, step by step
- Pick one bounded workflow, such as baggage policy Q&A or disruption drafts.
- Classify the passenger and operational data the workflow actually needs.
- Ground answers with embeddings and RAG over current policy documents.
- Build against the OpenAI-compatible endpoint and keep prompts and evals in version control.
- Choose the deployment tier, region and failover behaviour your obligations require.
- Pilot on historical disruption cases, measure handling time, then expand.
Try it yourself
Open the AI agent cost calculator →
Where an AI API fits in airlines
Airline operations generate policy-heavy text and disruption-driven spikes. The best first use cases are grounded in your own documents and measured by handling time:
- Disruption messaging: draft clear rebooking and compensation guidance from the current policy, with agent approval before it reaches passengers.
- Policy Q&A: answer agent and passenger questions from fare rules, baggage policy and conditions of carriage with citations.
- Operations knowledge search: surface the right procedure or bulletin from maintenance, ground and crew documentation.
- Cargo documentation: summarise and check air waybills and customs paperwork for missing fields, with a trained reviewer signing off.
- Contact-centre summaries: condense long case histories so agents start every conversation with context.
Security, privacy and data handling
Airlines process passenger name records, payment data and safety-related documentation, so the API layer needs the same discipline as any other passenger system:
- Minimise PNR and payment fields in prompts; pass reference numbers, not full profiles.
- Keep safety-critical and airworthiness decisions out of model scope entirely.
- Use per-service keys and RBAC so contact-centre, cargo and operations access stays separate.
- Enable audit export and retention controls to satisfy cross-border and records obligations.
Deployment options and model choice
Carrier IT estates are hybrid by nature, so keep passenger-facing assistants in a region that matches route privacy commitments. Plugsky exposes the same OpenAI-compatible endpoint whether the workload runs in shared cloud, your VPC, on-prem or air-gapped, and region-locked planes keep processing where your policy requires. Teams route to 30+ models through one API, with open-weight options when a site has no external connectivity, and port code with a base URL change. Chat, streaming, JSON mode, function calling, embeddings, RAG and agents are live; audio, images, moderation, files, batch, fine-tuning, assistants and the responses API are coming soon. plugsky-micro and plugsky-lite are free with no card, a 14-day full-access trial covers paid tiers, and the live pricing page lists current plans.
From pilot to production
Disruption is the stress test: a model that works in a demo but not during an irregular operation creates more work than it saves. Common mistakes:
- Designing for steady-state volume and ignoring disruption spikes.
- Letting generated passenger messages go out without agent or policy review.
- Passing full passenger profiles when a booking reference would do.
- Mixing contact-centre and operations workloads on one key with no attribution.
- Skipping a fallback path for when the API or the network is degraded.
Pilot on one disruption scenario with historical cases, measure draft acceptance and handling time, and rehearse a degraded mode before go-live. Failover, status monitoring and rollback are part of the product, not extras.
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 and sovereign options | 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 |
| Industry fit | Disruption messaging, policy Q&A, operations knowledge search, cargo documents | Generic API, you adapt it | You build every workflow |
Frequently asked questions
Can we keep our existing OpenAI SDK code?
Yes. Plugsky exposes an OpenAI-compatible API, so you change the base URL and model name and keep your SDK and prompt structure.
Is there a free plan?
Yes. The free plan includes two free models, plugsky-micro and plugsky-lite, and does not require a credit card.
How does pricing work?
Self-serve plans are flat monthly with unlimited fair-use usage; enterprise agreements cover residency, capacity and SLA terms. See the live pricing page for current plans.
Can we avoid sending full passenger profiles?
Yes, and you should. Reference numbers, route segments and policy text usually answer the question without transferring a full PNR.
What happens during disruption spikes?
Plan capacity and failover up front. Automatic upstream failover and the status page help, but the application still needs a degraded mode for agents.
Which endpoints are live today?
Chat completions, streaming, JSON mode, function calling, embeddings, RAG and agents are live. Audio, images, moderation, files, batch, fine-tuning, assistants and the responses API are coming soon.
Can the API touch safety-critical workflows?
No. Keep airworthiness, dispatch and safety decisions out of scope; use the API for documents, communications and knowledge retrieval with human review.