docs: add missing integration guides (streaming, tool-calling, anthropic-messages, embeddings, responses, audio, images, rerank, passthrough, errors-and-retries)

[moonming]

Background

Follow-up to #246 (closed by #249). The docs rebuild shipped one of the eleven planned integration pages at full quality, plus a baseline errors-and-retries.md:

• ✅ docs/integration/openai-compatible-api.md — full request-flow mermaid + endpoint surface + auth
• ✅ docs/integration/errors-and-retries.md — status-code table, error.type strings verified against ProxyError

The other ten pages exist but are placeholder-tier and need expansion. Per #240's "Remaining Work Breakdown", they are tracked here as separate doc refinements.

Scope

Expand each page to match the depth of openai-compatible-api.md (request shape, response shape, streaming behavior where applicable, provider compatibility, real curl + SDK example, anchor to the corresponding e2e test):

• docs/integration/streaming.md — SSE event ordering, abort behavior, partial-chunk semantics
• docs/integration/tool-calling.md — cross-provider tool-use translation; anchor to tests/e2e/src/cases/tools-cross-provider-e2e.test.ts
• docs/integration/anthropic-messages.md — full /v1/messages reference; OpenAI → Anthropic translation matrix
• docs/integration/embeddings.md — providers that implement it, dimension handling
• docs/integration/responses.md — OpenAI Responses API surface; current OpenAI-provider-only constraint
• docs/integration/audio.md — transcriptions / translations / speech; per-provider compatibility
• docs/integration/images.md — /v1/images/generations; provider compatibility
• docs/integration/rerank.md — Jina AI rerank surface (per feat(rerank): Jina AI provider support on /v1/rerank (#213 Phase 2) #229)
• docs/integration/passthrough.md — /passthrough/:provider/*rest semantics, auth boundary

Acceptance criteria

For each page:

• request body shape with every field documented in a table (type / required / description)
• response body shape with every observable field
• error envelope behavior for endpoint-specific failure modes
• at least one runnable curl example + at least one SDK example where the SDK supports the endpoint
• citation of the e2e test that pins the contract

References

• Parent rebuild PR: docs: rebuild customer-facing docs across overview, config, cloud, ops, reference, and tutorials #249
• Original integration-guides issue: docs: create client integration guides for OpenAI-compatible and Anthropic-style APIs #246

[coderabbitai]

🔗 Related PRs

#102 - docs: add docs about Introduction, Quick Start, Core Concepts, and Guides [merged]
api7/docs#1431 - docs: add Anthropic-OpenAI protocol conversion guide for ai-proxy plugin [merged]
api7/docs.apiseven.com#212 - docs: 新增 Anthropic 与 OpenAI 协议转换使用指南 [merged]
api7/contents#814 - Fix proxy path mismatch in Getting Started example: add proxy-rewrite to strip /api prefix [merged]
api7/api7ee-3-control-plane#2609 - docs(AGENTS): require OpenAPI sync and prefer integration tests [merged]

---

📝 Issue Planner

_{Check the box below or use the @coderabbitai plan command to generate an implementation plan and prompts that you can use with your favorite coding assistant.}

• Create Plan

---

🧪 Issue enrichment is currently in open beta.

You can configure auto-planning by selecting labels in the issue_enrichment configuration.

To disable automatic issue enrichment, add the following to your .coderabbit.yaml:

issue_enrichment:
  auto_enrich:
    enabled: false

💬 Have feedback or questions? Drop into our discord!