feat: add MiniMax provider support

[octo-patch]

Reason: add target provider/model to existing provider registry

• Register MiniMax-M3 and MiniMax-M2.7 with context, pricing, thinking, and input-modality metadata.
• Add global and China endpoint selection with both compatible base URL families retained in the registry.
• Wire provider selection, authentication routing, model display, cost/context handling, and bilingual setup documentation.

Checks:

• git diff --check
• Syntax-build all changed TypeScript files with Bun
• Target coverage assertions
• Provider registry assertions
• node package/cli.js --version
• Secret scan

Summary by CodeRabbit

• New Features
  • Added built-in MiniMax provider support with region-based endpoints, authentication, and Anthropic-compatible base URL overrides.
  • Enabled MiniMax model selection with model labels, context window, and supported thinking-mode options.
  • Preserve MiniMax routing settings across managed environments and teammate sessions.
  • Updated provider status to show the selected MiniMax region and endpoint.
• Bug Fixes
  • Disabled analytics, feedback actions, error reporting, and Anthropic preconnects when MiniMax is enabled.
• Documentation
  • Updated Quick Start and configuration guides, including region/Base URL tables and model enablement examples; clarified protocol-conversion guidance for OpenAI-only endpoints.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: ca4ecb8f-45ca-4f7d-874a-871779cae099

📥 Commits

Reviewing files that changed from the base of the PR and between 9967251 and f4cb081.

📒 Files selected for processing (3)

• README.en.md
• README.md
• src/utils/model/providers.ts

🚧 Files skipped from review as they are similar to previous changes (2)

• README.en.md
• README.md

---

📝 Walkthrough

Walkthrough

MiniMax is added as an API provider with registered models, regional endpoints, pricing, model capabilities, runtime routing, environment handling, status reporting, and README configuration guidance.

Changes

MiniMax provider integration

Layer / File(s)	Summary
Provider registry and model contracts src/utils/model/providers.ts, src/utils/model/configs.ts, src/utils/model/deprecation.ts, src/utils/modelCost.ts	Defines MiniMax models, regions, endpoints, provider selection, Claude-to-MiniMax mappings, retirement metadata, and pricing.
Model selection and capabilities src/utils/model/model.ts, src/utils/model/modelOptions.ts, src/utils/context.ts, src/utils/thinking.ts	Adds MiniMax model picker entries, display-name resolution, context-window lookup, and thinking capability detection.
Runtime routing and environment handling src/services/api/client.ts, src/utils/auth.ts, src/utils/apiPreconnect.ts, src/utils/log.ts, src/services/analytics/config.ts, src/commands/feedback/index.ts, src/utils/managedEnvConstants.ts, src/utils/swarm/spawnUtils.ts	Routes API clients to regional or overridden MiniMax endpoints and applies provider-specific authentication, telemetry, preconnect, managed-environment, command, and teammate-process behavior.
Status reporting and configuration documentation src/utils/status.tsx, README.en.md, README.md	Reports MiniMax region and Anthropic base URL details and documents environment setup, model selection, regional URLs, and endpoint overrides.

Estimated code review effort: 4 (Complex) | ~45 minutes

Sequence Diagram(s)

sequenceDiagram
  participant User
  participant Environment
  participant getAPIProvider
  participant APIClient
  participant AnthropicSDK
  User->>Environment: Set MiniMax provider and region variables
  Environment->>getAPIProvider: Read CLAUDE_CODE_USE_MINIMAX
  getAPIProvider-->>APIClient: Return minimax provider
  APIClient->>AnthropicSDK: Configure MiniMax Anthropic base URL
  User->>AnthropicSDK: Send model request
  AnthropicSDK-->>User: Return MiniMax response

Loading

🚥 Pre-merge checks | ✅ 4 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 40.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (4 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title accurately summarizes the main change: adding MiniMax provider support across the codebase.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 2

🧹 Nitpick comments (1)

src/utils/modelCost.ts (1)

90-108: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Prefer modelId-based lookup over array index access for MiniMax cost constants.

MINIMAX_M3_COST and MINIMAX_M27_COST are built using MINIMAX_MODELS[0] and MINIMAX_MODELS[1]. If the array is reordered, the wrong pricing would be silently associated with the wrong model — TypeScript won't catch it because both elements satisfy ModelCosts. The MODEL_COSTS keys at lines 147–148 would also point to the wrong cost object.

♻️ Proposed refactor: lookup by modelId

+const minimaxM3 = MINIMAX_MODELS.find(m => m.modelId === 'MiniMax-M3')!
+const minimaxM27 = MINIMAX_MODELS.find(m => m.modelId === 'MiniMax-M2.7')!
+
 const MINIMAX_M3_COST = {
-  inputTokens: MINIMAX_MODELS[0].pricingUsdPerMillionTokens.input,
-  outputTokens: MINIMAX_MODELS[0].pricingUsdPerMillionTokens.output,
+  inputTokens: minimaxM3.pricingUsdPerMillionTokens.input,
+  outputTokens: minimaxM3.pricingUsdPerMillionTokens.output,
   promptCacheWriteTokens:
-    MINIMAX_MODELS[0].pricingUsdPerMillionTokens.cacheWrite,
+    minimaxM3.pricingUsdPerMillionTokens.cacheWrite,
   promptCacheReadTokens:
-    MINIMAX_MODELS[0].pricingUsdPerMillionTokens.cacheRead,
+    minimaxM3.pricingUsdPerMillionTokens.cacheRead,
   webSearchRequests: 0,
 } as const satisfies ModelCosts

 const MINIMAX_M27_COST = {
-  inputTokens: MINIMAX_MODELS[1].pricingUsdPerMillionTokens.input,
-  outputTokens: MINIMAX_MODELS[1].pricingUsdPerMillionTokens.output,
+  inputTokens: minimaxM27.pricingUsdPerMillionTokens.input,
+  outputTokens: minimaxM27.pricingUsdPerMillionTokens.output,
   promptCacheWriteTokens:
-    MINIMAX_MODELS[1].pricingUsdPerMillionTokens.cacheWrite,
+    minimaxM27.pricingUsdPerMillionTokens.cacheWrite,
   promptCacheReadTokens:
-    MINIMAX_MODELS[1].pricingUsdPerMillionTokens.cacheRead,
+    minimaxM27.pricingUsdPerMillionTokens.cacheRead,
   webSearchRequests: 0,
 } as const satisfies ModelCosts

Also applies to: 147-148

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/utils/modelCost.ts` around lines 90 - 108, Replace positional
MINIMAX_MODELS[0] and MINIMAX_MODELS[1] access in MINIMAX_M3_COST and
MINIMAX_M27_COST with lookups by the corresponding modelId, preferably using a
shared helper that finds and validates the expected model. Update the
MODEL_COSTS entries to reference costs by explicit model IDs so reordering
MINIMAX_MODELS cannot associate incorrect pricing.

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@src/utils/status.tsx`:
- Around line 324-333: Update the MiniMax branch in the status-building logic to
ignore an empty ANTHROPIC_BASE_URL override by using a truthy fallback check
instead of nullish coalescing, matching neighboring provider branches and
preserving endpoint.anthropicBaseUrl when the environment value is blank.
- Line 373: Regenerate the inline source map at the end of src/utils/status.tsx
so its embedded sourcesContent and mappings match the current implementations,
especially buildAPIProviderProperties and related provider-label logic; use the
project’s standard TypeScript build or source-map generation workflow rather
than editing the base64 payload manually.

---

Nitpick comments:
In `@src/utils/modelCost.ts`:
- Around line 90-108: Replace positional MINIMAX_MODELS[0] and MINIMAX_MODELS[1]
access in MINIMAX_M3_COST and MINIMAX_M27_COST with lookups by the corresponding
modelId, preferably using a shared helper that finds and validates the expected
model. Update the MODEL_COSTS entries to reference costs by explicit model IDs
so reordering MINIMAX_MODELS cannot associate incorrect pricing.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro

Run ID: 0532de8e-e6fb-4f60-9a72-0d4d3ac50a2e

📥 Commits

Reviewing files that changed from the base of the PR and between e54c5af and 9967251.

📒 Files selected for processing (19)

• README.en.md
• README.md
• src/commands/feedback/index.ts
• src/services/analytics/config.ts
• src/services/api/client.ts
• src/utils/apiPreconnect.ts
• src/utils/auth.ts
• src/utils/context.ts
• src/utils/log.ts
• src/utils/managedEnvConstants.ts
• src/utils/model/configs.ts
• src/utils/model/deprecation.ts
• src/utils/model/model.ts
• src/utils/model/modelOptions.ts
• src/utils/model/providers.ts
• src/utils/modelCost.ts
• src/utils/status.tsx
• src/utils/swarm/spawnUtils.ts
• src/utils/thinking.ts

[coderabbitai]

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

Ignore empty ANTHROPIC_BASE_URL overrides.

With ANTHROPIC_BASE_URL="", nullish coalescing selects the empty string, so the CLI status consumer prints a blank MiniMax base URL instead of the selected endpoint. Neighboring provider branches already use truthy checks for this case.

-      value: process.env.ANTHROPIC_BASE_URL ?? endpoint.anthropicBaseUrl
+      value: process.env.ANTHROPIC_BASE_URL || endpoint.anthropicBaseUrl

📝 Committable suggestion

‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.

Suggested change

	} else if (apiProvider === 'minimax') {
	const endpoint = getMiniMaxEndpoint();
	properties.push({
	label: 'MiniMax region',
	value: endpoint.region
	});
	properties.push({
	label: 'MiniMax Anthropic base URL',
	value: process.env.ANTHROPIC_BASE_URL ?? endpoint.anthropicBaseUrl
	});
	} else if (apiProvider === 'minimax') {
	const endpoint = getMiniMaxEndpoint();
	properties.push({
	label: 'MiniMax region',
	value: endpoint.region
	});
	properties.push({
	label: 'MiniMax Anthropic base URL',
	value: process.env.ANTHROPIC_BASE_URL || endpoint.anthropicBaseUrl
	});

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/utils/status.tsx` around lines 324 - 333, Update the MiniMax branch in
the status-building logic to ignore an empty ANTHROPIC_BASE_URL override by
using a truthy fallback check instead of nullish coalescing, matching
neighboring provider branches and preserving endpoint.anthropicBaseUrl when the
environment value is blank.

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

python3 - <<'PY'
import base64, json, re, pathlib

path = pathlib.Path("src/utils/status.tsx")
text = path.read_text()

m = re.search(r'sourceMappingURL=data:application/json;charset=utf-8;base64,([A-Za-z0-9+/=]+)', text)
if not m:
    print("no inline source map found")
    raise SystemExit(0)

data = json.loads(base64.b64decode(m.group(1)).decode("utf-8"))
src = data["sourcesContent"][0]

checks = [
    ("getMiniMaxEndpoint import", "getMiniMaxEndpoint" in src, "getMiniMaxEndpoint" in text),
    ("MiniMax branch", "miniMax" in src or "MiniMax" in src, "miniMax" in text or "MiniMax" in text),
    ("pre-MiniMax provider labels", "API provider" in src, "API provider" in text),
]

print("sources:", data.get("sources"))
for name, in_src, in_file in checks:
    print(f"{name}: sourcesContent={in_src} file={in_file}")
print("sourcesContent matches file prefix:", text.startswith(src[:200]))
print("sourcesContent length:", len(src))
print("file length:", len(text))
PY

Repository: LING71671/Open-ClaudeCode

Length of output: 441

---

Regenerate the inline source map src/utils/status.tsx:373 — the embedded sourcesContent is stale and still reflects the old provider-label branch, so debugger and stack-trace mappings won’t line up with this file.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/utils/status.tsx` at line 373, Regenerate the inline source map at the
end of src/utils/status.tsx so its embedded sourcesContent and mappings match
the current implementations, especially buildAPIProviderProperties and related
provider-label logic; use the project’s standard TypeScript build or source-map
generation workflow rather than editing the base64 payload manually.