OrcaRouter provider support for Sentry MCP
维护者通常 1 天内回复
还没有人认领这个 Issue。
评估
- 难度
- 5/5
- 预计耗时
- 一周以上
- 新手友好度
- 30/100
- Issue 类型
- 功能
- 描述清晰度
- 需要澄清
- 活跃度
- 活跃
- 技术栈
- typescript
- 领域
- ai, backend-api-design
调研方向
从 packages/mcp-core/src/internal/agents/provider-factory.ts 和 issue 中描述的 AgentProviderType 注册开始。将提议的 OrcaRouter 条目与现有的 OpenRouter 配置进行比较，包括其 API-key 变量和可选的 base URL。在编写代码之前，维护者就 provider 的结构和符合披露要求的实现方向达成一致，即视为完成。
由索引模型根据 Issue 内容生成。
描述
OrcaRouter provider support for Sentry MCP
Hi Sentry MCP maintainers,
sentry-mcp wraps Sentry's error tracking and performance monitoring into 19 MCP tools built for human-in-the-loop coding agents like Cursor and Claude Code. I especially respect how you've handled the genuinely hard part: the natural-language tools (search_events, search_issues, search_issue_events) depend on an LLM provider, so provider selection is explicit and conflict-proof via EMBEDDED_AGENT_PROVIDER and per-provider API-key variables, backed by a small provider factory.
That design is also why adding one more option is useful rather than redundant: teams that standardize Sentry debugging across many developers, or people who self-host the stdio transport, get the AI search tools working against whichever model endpoint fits their stack. A dedicated OrcaRouter provider would slot alongside OpenRouter as a new explicit choice, without changing any existing configuration.
Proposal
I'd like to propose OrcaRouter as an optional embedded-agent provider for sentry-mcp. It would only add a new choice for the AI-powered search tools; it would not replace or alter OpenAI, Azure OpenAI, Anthropic, or OpenRouter support.
A few capabilities seem most relevant to your users:
- Many models through one endpoint — chat, reasoning, image, and video models from multiple upstream providers behind a single endpoint, so users get model variety for query translation without juggling several keys.
- Automatic routing and provider failover — if an upstream provider or model is down or rate-limited, requests route to an alternative, which matters when your search tools depend on a provider being reachable.
- Usage tracking and budgets — per-key usage records and spend caps fit the cost awareness visible in the repo (tool-definition token-cost tracking, the
token-costCI workflow).
Expected integration point
OrcaRouter exposes an OpenAI-compatible API and uses standard API-key authentication. Embedded-agent providers are registered through AgentProviderType and the factory in packages/mcp-core/src/internal/agents/provider-factory.ts, each with its own key variable (e.g. OPENROUTER_API_KEY). My assumption — open to your preference — is an orcarouter entry alongside openrouter with an ORCAROUTER_API_KEY variable, or documenting it as an OpenAI-compatible base URL. I haven't written any code; this issue is to agree on the shape first.
Ecosystem
OrcaRouter is already shipped by open-source tools, including developer-facing projects such as goose and OpenCode; more examples are at https://www.orcarouter.ai/built-with.
Disclosure
Transparently: OrcaRouter runs an optional open-source partner program through which approved OSS projects can receive a 5% revenue share from OrcaRouter usage attributed to their integration. Participation is not a precondition for support, and I'll follow this repository's disclosure and governance requirements either way.
I'm an engineer on the OrcaRouter team. Would an optional OrcaRouter provider be of interest to sentry-mcp? I'm happy to refine the design and submit an implementation PR once you've approved the direction.
- 主要语言
- TypeScript
- 星标
- 938
- 派生
- 166
- 平均合并
- 1 天 7 小时
- 30 天内合并 PR
- 156
环境准备
这个项目没有提供开发容器、Dockerfile 或贡献指南，环境需要你自己搭建：先看它的 README，通用步骤见我们的新手贡献指南。
从这里开始
- 先读完整个 Issue，再读项目的贡献指南。
- 在 Issue 下留言说明你要接手 —— 这能避免两个人做同样的事。
- Fork 仓库，在一个分支上完成修改。
- 提交 Pull Request，并在描述里引用这个 Issue 编号。
getsentry/toolkit 的其他 Issue
-
bug MCP
难度 2/5 1-3 小时 新手友好度 78/100
getsentry/toolkit#1600 · 1 条评论 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 72/100
getsentry/toolkit#1471 · 1 条评论 ·
维护者通常 1 天内回复
-
难度 4/5 3-5 天 新手友好度 52/100
getsentry/toolkit#1611 · 1 条评论 ·
维护者通常 1 天内回复
-
enhancement MCP
难度 4/5 1-2 天 新手友好度 55/100
getsentry/toolkit#1601 · 1 条评论 ·
维护者通常 1 天内回复
-
Use typed detector schemas from @sentry/api for metric monitors可能已有人在做 @MathurAditya724 于 1 天前认领。 未关闭
getsentry/toolkit#1595 · 1 条评论 · 已指派 1 人 ·
维护者通常 1 天内回复
查看 getsentry/toolkit 的全部 Issue
相似的 Issue
-
难度 2/5 1-3 小时 新手友好度 66/100
lyswhut/lx-music-desktop#3026 ·
-
area: ci type: test
难度 2/5 1-3 小时 新手友好度 65/100
schlessera/brain-kit#2769 ·
维护者通常 1 天内回复
-
Read a vector index's quantization back on Neo4j 2026.10, which reports vector.quantization.type未关闭
难度 2/5 1-3 小时 新手友好度 78/100
liberation-data/drivine#231 ·
维护者通常 1 天内回复
-
bug
难度 2/5 1-3 小时 新手友好度 65/100
alveusgg/alveusgg#2502 · 3 条评论 ·
维护者通常 1 天内回复
-
难度 2/5 1-3 小时 新手友好度 78/100
维护者通常 1 天内回复