JobSync is a self-hosted job search companion that takes on the tedious parts of a search - tracking applications, reviewing resumes, matching them against live postings, and drafting cover letters from a docked assistant panel. I especially like how the automated discovery runs for Greenhouse, Lever, and Ashby are cost-bounded: listings are pre-ranked locally, and only the top candidates are sent to the model.
That restraint is backed by a clean provider abstraction. ProviderType in src/lib/ai/providers.ts and the PROVIDER_REGISTRY in src/lib/ai/provider-registry.ts let each backend register an api-key entry with a models endpoint, and the README is upfront that models must support tool calling and structured output. For people who self-host, having a real choice of model backends without lock-in is part of the appeal.
Proposal
I'd like to propose adding OrcaRouter as an optional provider in Settings > AI Settings. It would be a purely additive entry alongside Ollama, OpenAI, DeepSeek, Gemini, and OpenRouter - nothing about the existing providers would change.
OrcaRouter exposes an OpenAI-compatible API and uses standard API-key authentication, so it fits JobSync's registry the same way OpenRouter does today: a cloud/api-key entry with an envVar, a modelsEndpoint, a factory in provider-registry.server.ts, and a models route mirroring src/app/api/ai/openrouter/models/route.ts. To be clear, this is describing the integration seam, not an implementation - nothing is written or tested yet.
Why it would be useful for JobSync specifically:
- One key, many models - chat, reasoning, image, and video models through a single endpoint, so a user can pick the best model per task instead of juggling several provider keys.
- Automatic routing and provider failover - if one upstream is down or rate-limited, requests move to a healthy provider; that matters for scheduled job-discovery matches you would rather not babysit.
- Usage tracking and budgets - per-key usage and spend caps pair naturally with JobSync's cost-bounded discovery design.
OrcaRouter is already integrated with open-source projects including Dify, RAGFlow, and OpenCode. I'm an engineer on the OrcaRouter team.
Full transparency: OrcaRouter has an optional open-source partner program through which approved OSS projects can receive a 5% revenue share from OrcaRouter usage attributed to their integration. Taking part is entirely optional and not a prerequisite for the integration; if you'd rather not, we can simply skip it, and I'm happy to follow any disclosure or governance requirements JobSync has.
If this sounds aligned with JobSync's direction, I'd welcome your thoughts - and with your go-ahead, I'd be glad to open a PR adding the provider entry and models route. More context: https://www.orcarouter.ai/built-with
JobSync is a self-hosted job search companion that takes on the tedious parts of a search - tracking applications, reviewing resumes, matching them against live postings, and drafting cover letters from a docked assistant panel. I especially like how the automated discovery runs for Greenhouse, Lever, and Ashby are cost-bounded: listings are pre-ranked locally, and only the top candidates are sent to the model.
That restraint is backed by a clean provider abstraction.
ProviderTypeinsrc/lib/ai/providers.tsand thePROVIDER_REGISTRYinsrc/lib/ai/provider-registry.tslet each backend register an api-key entry with a models endpoint, and the README is upfront that models must support tool calling and structured output. For people who self-host, having a real choice of model backends without lock-in is part of the appeal.Proposal
I'd like to propose adding OrcaRouter as an optional provider in Settings > AI Settings. It would be a purely additive entry alongside Ollama, OpenAI, DeepSeek, Gemini, and OpenRouter - nothing about the existing providers would change.
OrcaRouter exposes an OpenAI-compatible API and uses standard API-key authentication, so it fits JobSync's registry the same way OpenRouter does today: a
cloud/api-keyentry with anenvVar, amodelsEndpoint, a factory inprovider-registry.server.ts, and a models route mirroringsrc/app/api/ai/openrouter/models/route.ts. To be clear, this is describing the integration seam, not an implementation - nothing is written or tested yet.Why it would be useful for JobSync specifically:
OrcaRouter is already integrated with open-source projects including Dify, RAGFlow, and OpenCode. I'm an engineer on the OrcaRouter team.
Full transparency: OrcaRouter has an optional open-source partner program through which approved OSS projects can receive a 5% revenue share from OrcaRouter usage attributed to their integration. Taking part is entirely optional and not a prerequisite for the integration; if you'd rather not, we can simply skip it, and I'm happy to follow any disclosure or governance requirements JobSync has.
If this sounds aligned with JobSync's direction, I'd welcome your thoughts - and with your go-ahead, I'd be glad to open a PR adding the provider entry and models route. More context: https://www.orcarouter.ai/built-with