# API Atlas — OpenAI + Anthropic + xAI + Gemini — Documentation index

Snapshot date: 2026-09-18/19 UTC (OpenAI + Anthropic on 09-18, xAI + Gemini on 09-19; all sources hashed, all live calls logged). Status legend: LIVE_VERIFIED called successfully with our keys · LIVE_DISCOVERED seen live, weakly documented · DOCUMENTED in current official docs · BETA / PREVIEW · LEGACY / DEPRECATED / RETIRED · ACCOUNT_RESTRICTED documented but refused for our keys · UNVERIFIED · FAILED_VERIFICATION. Machine-readable twins of everything below live in generated/ (see repository and data model).

# Cross-cutting

Topic	Page
FAQ: answering "which model / which JSON / which events / which header"	faq.md
Endpoint catalogue (4 providers, 870 endpoints)	endpoints/index.md · by status
Comparisons OpenAI ↔ Anthropic ↔ xAI ↔ Gemini	comparisons/index.md · models · features · state management · tool execution · streaming · agents platforms · pricing · caching & reasoning · realtime & media · Responses vs Chat Completions
Errors	errors/openai.md · errors/anthropic.md · errors/gemini.md · xai/authentication-headers-errors.md
Security	security/README.md (14 topic pages: keys, prompt/tool injection, MCP, browsing, computer use, code execution, uploads/SSRF, command/path injection, secret leakage, untrusted outputs, schema validation, allowlists, webhooks, least privilege)
Architecture	resilience · multi-provider abstraction · feature detection · capability graph · streaming patterns · repository & data model

# Models

	OpenAI	Anthropic	xAI	Gemini
Catalogue (Model × Capability / Endpoint / Tool matrices)	models/openai-models.md	models/anthropic-models.md	models/xai-models.md	models/gemini-models.md
Families & guidance	openai/model-families.md	(in catalogue)	(in catalogue)	(in catalogue)
Pricing	openai/pricing.md	anthropic/pricing.md	xai/pricing.md	gemini/pricing.md
Rate limits	openai/rate-limits.md	anthropic/rate-limits.md · service tiers	xai/rate-limits.md	gemini/rate-limits.md
Deprecations	openai/deprecations.md	anthropic/deprecations.md	xai/deprecations-and-release-notes.md	gemini/deprecations-and-changelog.md
Auth / headers / versioning	openai/headers.md · authentication	anthropic/versioning-and-beta-headers.md · headers · release notes digest	xai/authentication-headers-errors.md	gemini/authentication-headers-versions.md
Regions / residency / clouds	openai/data-residency-and-regions.md	anthropic/regions-and-ips.md · cloud providers	(in auth page: us./eu-west-1. endpoints)	gemini/vertex-vs-gemini-api.md
SDKs	openai/sdks.md	anthropic/sdks.md	xai/sdks.md · gRPC	gemini/sdks.md · OpenAI compatibility

# OpenAI

Area	Pages
Core generation	responses.md · chat-completions.md · completions-legacy.md · conversations.md · streaming-events.md · structured-outputs.md · reasoning.md · prompt-caching.md · multimodal-input.md
Tools	tools/openai/index.md · function calling · custom tools · web search · file search · code interpreter · computer use · image generation · MCP & connectors · shell · apply_patch · skills · tool search · deep research · tool-loop.md · containers.md · skills-api.md
Agents	agents-api.md · environments & vaults · chatkit.md · workspace-agents.md · assistants-retired.md · agents-sdk.md
Realtime / voice / audio	realtime.md · realtime-events.md · live.md · audio.md · voice-agents.md
Media	images.md · video.md · embeddings.md · moderation.md · safety.md
Platform	files-and-uploads.md · vector-stores.md · batch.md · fine-tuning.md · graders.md · evals.md
Administration & ops	admin-api.md · webhooks.md · authentication-and-keys.md · headers.md · sdks.md

# Anthropic

Area	Pages
Core generation	messages-api.md · content-blocks.md · streaming.md · token-counting.md · stop-reasons.md · text-completions-legacy.md
Advanced generation	prompt-caching.md · thinking.md · effort.md · fast-mode.md · context-management.md · structured-outputs.md · citations.md · vision-and-documents.md · files-api.md
Tools	tools/anthropic/index.md · custom tools & tool_choice · tool-use loop · fine-grained streaming · computer use · text editor · bash · memory · web search · web fetch · code execution · programmatic tool calling · tool search · MCP connector · advisor · browser use
Agents & platform	managed-agents.md · skills-api.md · message-batches.md
Administration & ops	admin-api.md · claude-code-analytics.md · compliance-and-iam.md · headers.md · sdks.md · cli.md · cloud-providers.md

# xAI (Grok)

Entry point: xai/index.md

Area	Pages
Core generation	chat-completions.md · responses.md (agentic server-side tools) · messages-compat.md (Anthropic-compatible) · streaming.md · structured-outputs.md · reasoning.md · prompt-caching.md · deferred-completions.md · legacy-completions.md
Tools	tools/xai/index.md · function calling · web search · x search · code execution · collections search · MCP · image understanding · tool-loop.md
Platform	files.md · collections.md · batches.md · skills-api.md · management-api.md
Media & voice	images.md · videos.md · voice.md (realtime WebSocket, TTS, STT)
Products	grok-build.md · grok-apps-and-integrations.md

# Gemini (Google)

Entry point: gemini/index.md

Area	Pages
Core generation	generate-content.md · streaming.md · structured-outputs.md · thinking.md · context-caching.md · safety.md · multimodal-input.md · token-counting.md · embeddings.md · files.md · legacy-palm-methods.md
Tools	tools/gemini/index.md · function calling · Google Search grounding · Google Maps grounding · URL context · code execution · computer use · File Search · MCP · tool-loop.md
Agents & realtime	interactions-api.md (Interactions, Deep Research, environments) · live-api.md · live-events.md
Media	image-generation.md · video-generation.md · music-generation.md · speech-generation.md · transcription.md
Platform	batch.md · tuning.md · long-running-operations.md

# Machine-readable outputs (`generated/`)

File	Content
`models.json` / `models.csv`	396 model records (230 OpenAI, 36 Anthropic, 33 xAI, 97 Gemini — incl. snapshots/aliases/retired)
`endpoints.json` / `.csv`	870 endpoint records (359 OpenAI, 282 Anthropic, ~110 xAI, ~120 Gemini)
`parameters.json` / `.csv`	7,862 parameter records (dotted paths, types, enums, model compatibility, beta headers)
`tools.json` / `.csv`	69 tool records with exact `type` strings (19 OpenAI, 27 Anthropic, 12 xAI, 11 Gemini)
`streaming-events.json`	506 events (Responses, Responses WebSocket, Chat, Realtime, Live, transcription, images, Agents, Messages, Managed Agents, xAI voice/inference, Gemini Live/Interactions/Lyria)
`errors.json`, `headers.json`, `pricing.json`/`.csv`, `deprecations.json`, `rate-limits.json`, `objects.json`, `sdks.json`, `webhook-events.json`, `audit-log-events.json`, `status-lifecycles.json`	see names
`compatibility/`	model × capability, model × endpoint, model × tool, tool × model, Anthropic feature × model, feature × platform (clouds), model × platform ids, cross-provider feature matrix
`capability-graph.{json,mmd,dot}`	Provider → API family → Endpoint → Model → Capability → Tool graph
`examples-manifest.json`	every example with its verification status
`build-report.json`, `index.json`	provenance, counts, conflicts

# Reports

final-report.md · coverage.json · changes.md (re-run diff) · link-check.json · live-requests.jsonl (every live call, sanitized).