bf16.ai Remote MCP
Server MCP jarak jauh yang menyediakan penemuan model LLM tanpa sensor dengan konteks 1M, perbandingan harga, dan panduan penyiapan agen coding untuk Claude Code, Codex, dan OpenCode melalui api.bf16.ai/mcp.
Dokumentasi
bf16.ai Protocol & Gateway Reference
bf16.ai Developer API Documentation
Point any OpenAI SDK, Anthropic SDK, or coding agent at https://api.bf16.ai using a sk-bf16-v1-... Bearer key.
1. Supported Wire Endpoints
| Method & Path | Auth Header | Protocol Contract |
|---|---|---|
| POST https://api.bf16.ai/v1/chat/completions | Authorization: Bearer sk-bf16-v1-... | OpenAI Chat Completions (SSE streaming, reasoning_content, tools, JSON schema). |
| POST https://api.bf16.ai/v1/responses | Authorization: Bearer sk-bf16-v1-... | Stateless OpenAI Responses API (store: false enforced for Zero-Prompt Retention). |
| POST https://api.bf16.ai/v1/messages | x-api-key or Bearer sk-bf16-v1-... | Native Anthropic Messages wire format with HMAC-signed thinking blocks and tool_use translation. |
| GET https://api.bf16.ai/v1/models | Optional / Bearer | Live catalog metadata, 1M context limits, and per-token USD rates. |
2. Optional Control & Telemetry Headers
X-BF16-Max-Price-Per-M-Output: 2.00— Rejects routing to any endpoint whose output rate exceeds your USD ceiling.X-BF16-Include-Cost: true— Appends settled USD cost in response headers (X-BF16-Cost-USD) and final SSE usage frame.Idempotency-Key: <uuid>— Prevents duplicate non-streaming charges across network retries.
Rate limits
Limits apply per account, across all your API keys and models. They rise with your total top-ups, the moment a payment lands.
| Plan | Total top-ups | Concurrent requests | Requests per minute | Max input per request |
|---|---|---|---|---|
| Free | No payment | 3 | 180 | Full context |
| Tier 1 | $10+ | 10 | 600 | Full context |
| Tier 2 | $100+ | 30 | 1,800 | Full context |
| Tier 3 | $500+ | 50 | 3,000 | Full context |
| Tier 4 | $2,000+ | 100 | 6,000 | Full context |
- Refunds and chargebacks lower your total; promotional credit does not count.
- A request counts toward concurrency from the moment it is accepted until its response finishes.
- Purchased credit never expires.
When you hit a limit
400
Context window overflow (context_length_exceeded) or submitted array above the route ceiling (too_many_messages, default 100,000 submitted entries on messages / input, exposed as bf16.max_messages on GET /v1/models; max 1,024 tools). Compact the session history and retry.
429
Your plan’s concurrent-request or per-minute limit is in use. Retry after the time in the Retry-After header, or top up to raise the limit.
503
The model is busy or briefly unavailable. Retry after the time in the Retry-After header.
Responses carry x-ratelimit-limit-requests, x-ratelimit-remaining-requests and x-ratelimit-reset-requests. GET /v1/me returns your plan, balance and concurrency in use.
Top up Need more? For higher limits, tell us about your workload. Contact us
3. Zero-Prompt Retention Guarantee
bf16.ai streams tokens directly through memory without persisting request bodies, system prompts, or model completions to disk or databases. Only token counts, HTTP status codes, and nano-USD ledger postings are logged.