Axum AI Starter

- Axum AI Starter v1.0.1 source archive

- SHA-256 integrity digest for the delivered release

- Rust/Tokio/Axum backend

- PostgreSQL + SQLx migrations and repositories

- OpenRouter JSON chat integration

- Application-level SSE streaming

- Persistence-safe JSON and streaming Retry endpoints

- Typed public error contract and request-ID correlation

- Timeout and conservative pre-acceptance retry handling

- Streaming cancellation/disconnect persistence

- Next.js + React + TypeScript reference frontend

- Reusable JSON API client, SSE parser/client, and shared TypeScript types

- Multi-turn state, Stop generating, Retry, New chat, and error/loading UI

- Mock-provider tests with no paid AI calls

- PostgreSQL integration and fresh-database verification tooling

- GitHub Actions CI and release-quality gates

- Backend Dockerfile and Docker Compose local PostgreSQL setup

- Quickstart and architecture documentation

- HTTP API and streaming documentation

- Database and testing guides

- Deployment and customization guides

- Troubleshooting guide

- Security policy

- Support/update policy

- Production-readiness checklist

- Commercial source license

### You may

- use Axum AI Starter in paid commercial products;

- use it in non-commercial products;

- modify the source for your derived applications;

- deploy applications built from it;

- use it for client projects;

- deliver a genuine derived application to a client, including source when reasonably necessary for that client to use or maintain the derived application;

- allow necessary employees or contractors to work with the starter for permitted projects under the access restrictions in the license.

### You may not

- resell Axum AI Starter itself;

- redistribute the starter as a standalone/reusable source-code product;

- publish the starter source publicly;

- give unrelated third parties reusable copies;

- repackage it as a competing starter, boilerplate, template, source-code bundle, or substantially similar reusable product.

Third-party packages, services, APIs, models, and other dependencies remain subject to their own licenses and terms.

**This summary is for convenience only and does not replace `LICENSE.md`. `LICENSE.md` contains the controlling terms.**

### Normal support includes

- reasonable setup questions;

- documentation errors or broken instructions;

- reproducible defects in the unmodified or minimally modified starter;

- clarification of the documented JSON, SSE, persistence, configuration, testing, and frontend contracts;

- diagnostic guidance to help determine whether a problem is in the starter or an external environment/dependency.

### Normal support does not include

- 24/7 support or a fixed SLA;

- unlimited consulting;

- building your complete application;

- custom feature implementation;

- custom cloud/deployment operations;

- authentication or authorization integration;

- RAG, embeddings, agents, or tool calling;

- billing/subscription integration;

- additional AI-provider integrations;

- unlimited debugging of unrelated buyer modifications;

- operating or incident-managing your production application.

**Support channel:** `[email protected]`

### Updates

A standard Axum AI Starter v1 purchase includes updates that are actually released within the **v1.x** line at no additional license fee.

There is no promised minimum release cadence, minimum number of releases, lifetime support, or lifetime updates.

Future major versions, including **v2.x**, are not included unless a specific sale offer explicitly says otherwise.

### 1. What exactly do I receive?

You receive the versioned Axum AI Starter source package for the purchased release, its SHA-256 integrity digest, the Rust/Axum backend, PostgreSQL/SQLx persistence layer, OpenRouter integration, JSON and SSE chat paths, Next.js/TypeScript reference frontend, tests, CI configuration, Docker/local infrastructure assets, technical documentation, commercial license, security policy, and support/update policy.

For the current release, the buyer archive is:

`axum-ai-starter-v1.0.1-source-eba4b6d25f4e.tar.gz`

### 2. Is this a complete SaaS boilerplate?

No.

Axum AI Starter is a focused AI application foundation. It intentionally does not include production authentication, billing/subscriptions, user/team administration, RAG, agents/tool calling, multi-provider routing, Kubernetes/Terraform, or complete SaaS infrastructure.

### 3. Is it production-ready?

No.

It is **production-minded, not production-ready**. It includes deliberate error, persistence, streaming, cancellation, testing, and operational boundaries, but your application still needs production controls appropriate to its environment.

### 4. Is authentication included?

No.

Production authentication and authorization are intentionally outside v1. The documentation identifies where application-specific auth can be introduced, but the implementation is your responsibility.

### 5. Is billing included?

No.

Billing and subscriptions are intentionally excluded from v1.

### 6. Does it include RAG?

No.

RAG, embeddings, vector databases, and document ingestion are not included.

### 7. Can I change the OpenRouter model?

Yes.

The backend uses the server-side `OPENROUTER_MODEL` configuration value. The model is intentionally backend-authoritative rather than selected by the browser.

### 8. Can I replace OpenRouter?

Yes, by modifying the project.

OpenRouter is the implemented v1 provider. The architecture and customization documentation identify the provider boundary so you can replace it, but the starter does not claim a built-in multi-provider plug-and-play router.

### 9. Can I use it commercially?

Yes, subject to `LICENSE.md`.

The license permits lawful commercial and non-commercial derived applications.

### 10. Can I use it for client projects?

Yes, subject to `LICENSE.md`.

Client projects are explicitly permitted.

### 11. Can a client receive the derived application's source?

Yes, when reasonably necessary for the client to use or maintain the genuine derived application, subject to the license restrictions.

The client is not given a separate right to extract and redistribute Axum AI Starter as a standalone or competing reusable starter.

### 12. Can I redistribute or resell the starter?

No.

You may not resell, publicly publish, give away, or redistribute Axum AI Starter itself as a reusable source product, starter, template, boilerplate, code bundle, or competing source-code product.

### 13. Do I need an OpenRouter API key?

For normal provider-backed runtime use, yes.

The backend requires non-empty `OPENROUTER_API_KEY` and `OPENROUTER_MODEL` configuration. Keep the API key server-side.

### 14. Do automated tests consume paid AI credits?

No.

Automated provider tests use localhost mock servers and do not require a real OpenRouter key, paid AI credits, or public provider access.

### 15. Is PostgreSQL required?

For the starter as shipped, yes.

Persistence is a core part of the implementation and normal backend startup connects to PostgreSQL. You can redesign or replace persistence in a derived application, but that is a project modification rather than the out-of-box setup.

### 16. How does SSE work at a high level?

After the pre-stream database/provider acceptance boundary succeeds, the backend returns an application-level Server-Sent Events stream.

A successful stream is:

`meta -> delta* -> done`

A handled post-acceptance failure is:

`meta -> delta* -> error`

The backend persists terminal state and does not transparently restart an accepted stream after partial output may already have reached the client.

### 17. What does Stop generating do?

In the reference frontend, Stop generating aborts an active accepted SSE request.

The closed downstream connection becomes the backend cancellation signal. Provider consumption stops and the assistant attempt is persisted as `cancelled`, retaining non-empty partial output when appropriate.

### 18. How does persistence-safe Retry work?

The Retry endpoint accepts the conversation identity rather than asking the browser to resend the original user text.

For an eligible failed/cancelled attempt, the backend reuses the existing persisted user turn and creates a new assistant attempt. This prevents the common mistake of persisting the same user message again on Retry.

### 19. Is Docker required?

Not inherently.

Docker Compose is provided for a predictable local PostgreSQL workflow, and the repository includes a backend Dockerfile. You can run PostgreSQL and the application through other infrastructure if you configure the required environment correctly.

### 20. What deployment support is included?

The repository includes deployment documentation covering release builds, the provided backend Dockerfile, environment configuration, PostgreSQL migrations, process supervision considerations, browser/backend CORS, TLS/reverse-proxy boundaries, SSE buffering/timeouts, health checks, graceful shutdown, release discipline, and rollback considerations.

It does not include a production orchestrator, Kubernetes/Terraform, managed cloud resources, a secret-management platform, or custom deployment work as part of the product price.

### 21. What normal support is included?

Normal support covers reasonable setup/documentation questions, reproducible starter-level defects, clarification of documented contracts, and enough diagnostic guidance to separate starter problems from external environment/provider issues.

It does not include unlimited custom development, production operations, or a fixed support SLA.

### 22. Are v1.x updates included?

Yes.

A standard v1 purchase includes updates that Abbaas Labs actually releases within the `v1.x` line at no additional license fee.

No minimum update cadence or minimum number of future releases is promised.

### 23. Is v2 included?

No, not by default.

Future major versions such as `v2.x` are separate unless a specific sale offer explicitly states otherwise.

### 24. Can I hire you for customization?

You can contact Abbaas Labs to discuss separate paid customization work.

Possible custom work may include authentication, custom AI workflows, schema changes, frontend work, RAG, additional providers, or deployment help. Availability, scope, price, and delivery terms are agreed separately and are not included in the starter purchase.

Axum AI Starter v1.0.1 — Delivery & Setup Notes

Delivered files:

- axum-ai-starter-v1.0.1-source-eba4b6d25f4e.tar.gz

- axum-ai-starter-v1.0.1-source-eba4b6d25f4e.tar.gz.sha256

Expected SHA-256:

306ca023255385eb620956a99104278f3505dcd8d5aba65fc34bb5dc27fb7684

After purchase, verify the archive with:

sha256sum -c axum-ai-starter-v1.0.1-source-eba4b6d25f4e.tar.gz.sha256

Expected result:

axum-ai-starter-v1.0.1-source-eba4b6d25f4e.tar.gz: OK

Getting started:

1. Verify the archive.

2. Extract the source package.

3. Read README.md.

4. Follow docs/quickstart.md.

Support:

[email protected]

Important:

Keep your OpenRouter API key server-side.

Never commit a real .env file or API key.

Never place the OpenRouter secret in a NEXT_PUBLIC_* frontend variable.

Commercial use and redistribution rights are governed by the LICENSE.md included with the delivered package. Store-wide refund, payment, and delivery terms apply separately.