Key facts
| API compatibility | OpenAI-compatible chat, embeddings and function calling |
| Use cases | Coding assistance, codebase Q&A, ticket triage, test drafting, review summaries |
| Model choice | 30+ models including coding-focused aliases |
| Data controls | Scoped API keys, RBAC, SSO, audit logs and region selection |
| Deployment | Plugsky cloud, your VPC, on-prem and air-gapped options |
| 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 |
| Product status | Chat, streaming, JSON mode, function calling, embeddings, RAG and agents are live |
TL;DR
- Answer questions across docs and code with retrieval and citations.
- Draft tests and review summaries; keep a reviewer accountable for merges.
- Keep source code in your network with VPC, on-prem or air-gapped deployment.
- Route cheap tasks to small models and reserve frontier models for hard reasoning.
- Freeze model versions for CI-facing features so behaviour is stable.
How it works, step by step
- Pick a bounded workflow such as doc Q&A or pull-request summary generation.
- Index repositories, ADRs and runbooks with structure-aware chunks and stable ids.
- Expose retrieval as a tool so the model answers from retrieved context, not memory.
- Scope keys per service and environment, and enable audit logging before CI use.
- Pin model versions for anything that comments on pull requests or gates a pipeline.
- Measure review time and test coverage change, then expand to the next workflow.
Try it yourself
Open the best model for coding selector →
Where an AI API fits in a software team
Engineering work has several high-volume text tasks that suit narrow automation:
- Codebase Q&A: answer questions across repositories, ADRs and runbooks with links to the source.
- Test drafting: generate edge-case tests for a function or endpoint for a developer to review.
- Review summaries: summarise a pull request and flag missing tests or risky changes.
- Ticket triage: classify, route and de-duplicate incoming bug reports.
Keep a human owner for merges, releases and anything that changes production behaviour.
Protecting source code and IP
Source code is the asset most teams want to keep inside their boundary. Choose the deployment that matches your policy: public cloud for public documentation, your VPC for internal code, on-prem for strict control, and air-gapped where nothing may leave. Within any mode, use scoped API keys per service and environment, RBAC and SSO for engineers, and audit logs for attribution. Avoid sending secrets, tokens and customer data in prompts, and add a prompt-side filter for credential patterns. Plugsky provides the controls; your engineering policy defines what may be sent.
Model routing and cost control
Engineering workloads have very different difficulty levels, so routing matters. Use small models for commit-message drafting, label classification and short doc answers; use coding-tuned models for generation; reserve frontier reasoning models for architecture questions and hard debugging. Because Plugsky exposes 30+ models behind one compatible API, switching is a model-name change and A/B tests are cheap. Self-serve pricing is flat monthly with fair-use usage, which keeps CI-adjacent workloads from producing unpredictable bills. Freeze versions for anything that runs unattended.
From prototype to team rollout
Prototype with plugsky-coder or plugsky-lite on one repository, then use the 14-day full-access trial to compare coding models on your own codebase rather than public benchmarks. Track suggestion acceptance, review minutes saved, test pass rate and false-positive flags. Keep prompts in the repository, treat them as code, and require review for any prompt used in CI. Roll out team by team, and re-run your evaluation set whenever a model or retrieval chunking strategy changes.
Honest comparison
| Capability | Plugsky | Typical per-token API | Building in-house |
|---|---|---|---|
| API compatibility | OpenAI-compatible chat, embeddings and tools | Usually compatible | Full rewrite |
| Code privacy | Cloud, VPC, on-prem and air-gapped options | Mostly cloud-only | You control fully |
| Model choice | 30+ models including coding aliases | Varies by provider | You host every model |
| Pricing | Flat monthly self-serve, fair-use usage | Per-token, harder to forecast | GPU plus operations cost |
| Access control | Scoped keys, RBAC, SSO and audit logs | Varies by provider | You build all of it |
Frequently asked questions
Can we keep our existing OpenAI SDK code?
Yes. Plugsky is OpenAI-compatible, so existing tools, SDKs and eval harnesses keep working after 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 documentation or code Q&A prototype.
Will our source code leave our network?
Only if you choose a cloud deployment. VPC, on-prem and air-gapped options are available so code can stay inside your boundary.
Which model is best for coding?
Test coding-focused aliases against your own repository and tests. Benchmarks are a starting point; your evaluation set is the deciding evidence.
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.
Can we use it in CI?
Yes, with scoped keys, pinned model versions and audit logging, and with a human reviewer for anything that gates a merge or release.