Inferenesia Cloud

Gateway service

Uptime

Grok free · tokens left 24h

Grok free accounts

Own accounts

API keys

Strategy

Latency 1h

Gateway & pool

Proxy pool

Failing proxies

Filters / RTK

Warmup

PID / port

Error rate 1h

Account pools

Grok free

Own (yours)

Combined

Grok quota 24h

Providers

Total requests

today

Pool capacity 24h

Used 24h

rolling window

Tokens left 24h

Daily requests

Requests per day (7 hari)

Daily total tokens

Total tokens used per day (7 hari)

#	Model	Provider	Requests	Tokens

#	Username	Role	Requests	Tokens

No providers configured

Total

Active

Expired

No credit

Banned

Unprobed

Inactive

off, not banned

Your accounts

Own active

Own total

	#	Provider	Identity	Status	Quota	Expires	Last refresh	Requests	Source	Created	Actions

selected

total · page /

Per page

Total

—

Succeeded

—

Failed

—

Success rate

—

By error class

Class	Count

By provider

Provider	OK	Fail	Total

Hourly activity

Hour	OK	Fail	Total

Warmup status

Interval —

State —

How it works

Warmup probes setiap akun aktif: kirim request kecil ke upstream. Jika 402 → mark no_credits. Jika 401/403 → refresh token. Jika banned → mark inactive.

Time	Username	Account	Provider	Model	Prompt	Completion	Total	Latency	Status	Stream

Showing latest · matched ·

Usage statistic

Total request & token per periode (bulan / tahun)

Total requests

Total tokens

Avg tokens / request

period average

Data points

Requests trend

Line chart — full width

Tokens trend

Total tokens over period — full width

console · reverse chronological feed

Request Details

Tips: kosongkan replacement untuk hapus string. Regex contoh: \bsecret\b

#	Pattern	Replacement	Type	Active

Account strategy

Cara membagikan load ke akun provider (scope: API key / global)

Sticky RR limit request per akun sebelum rotate

Token saver

Kurangi token masuk/keluar otomatis di setiap request

RTK — compress tool output

Lipat output git/grep/ls/tree/find yang panjang (hemat 60–90% input). Global default: aktif.

Min chars

Min turns

Saran global: 4000 chars · 2 turns — chat pendek tidak disentuh; agent multi-turn + tool dump panjang dikompres.

Caveman — terse output

System prompt singkat agar jawaban lebih pendek (~65% fewer tokens)

Ponytail — YAGNI bias

Dorong model ke kode minimal, reuse stdlib, avoid bloat

Headroom — context compress

Kompres context lewat service eksternal sebelum request

Routing & filters

Word filters

Swap/restore blocked words

Outbound proxy pool

Kirim upstream lewat proxy pool

Warmup schedule

Interval

5m60m120m

Cek kesehatan akun + refresh token secara berkala

CodeBuddy CN credits

Refresh interval

15m60m360m

Scheduler fetch meter credits untuk akun CBCN aktif (progress bar quota di Accounts)

API rate limit (RPM)

Default user RPM

Requests/minute per API key owned by role user

Admin keys = unlimited. Set 0 to disable RPM for everyone. Per-key daily limit (API Keys page) still applies separately.

Last health check

Format single: http://user:pass@host:port · bulk: 1 baris per proxy, opsional url|label|country

	#	URL	Label	Country	Active	Status	Fails

selected

#	Username	Role	RPM	Active	Created

Loading user detail…

Your AI usage overview

Total requests

Today requests

Input tokens

Output tokens

Total tokens

Chat sessions

User info

Username

RoleAdminUser

StatusActiveInactive

Created

RPM limit

blank = global default · 0 = unlimited · N = custom

Admin keys are always unlimited RPM.

API key · 1 user = 1 key

No API key

Status Active Inactive

Requests

Tokens

Last used

Created

Regenerate invalidates the old secret immediately. Update OpenCode / clients after.

Recent request log

Time	Model	Tokens	Latency	Status
No requests yet

API key & quick start

Key aktif dipakai di semua contoh di halaman ini

Base URL

https://inferenesia.cloud/v1

Your API key

Free model: grok-4.5 · Auth header: Bearer <key>

Models

Daftar model yang tersedia di gateway

GET /v1/models

#	Model id	Provider	Stream	Actions
			required optional

Loading models…

• Provider key: grok / codebuddy-cn · label: Grok / CodeBuddy CN

• Prefix opsional: inferenesia/… (legacy temp-ai/…), grok/…, cbc/…, codebuddy-cn/…

• Own pool: model:own

bash · list models

Authentication

Semua request /v1/* pakai API key aktif sesimu:

http · header

Authorization: Bearer

Endpoints

Method	Path	Description
GET	/v1/models	Daftar model tersedia
POST	/v1/chat/completions	Chat completion (JSON / SSE stream)
POST	/anthropic/v1/messages	Anthropic Messages wire (Claude models · x-api-key / Bearer)
POST	/v1/images/generations	Text-to-image dan image-editing (kirim image[])
GET	/health	Health check (no auth)

Chat Completions

POST /v1/chat/completions

Default system prompt: kalau messages[] tidak berisi entry role: system, router otomatis prepend "You are a helpful AI assistant." supaya upstream terima. Kirim system message sendiri → dipakai verbatim, default TIDAK di-inject.

json · request body

Contoh curl (JSON)

bash · non-stream

Contoh curl (stream)

bash · SSE stream

Contoh Python (OpenAI SDK)

python

Vision (image input)

Kirim gambar lewat messages[].content multimodal. Model vision: glm-5v-turbo (CodeBuddy CN). Grok chat juga menerima image_url untuk analisis gambar.

bash · vision chat

python · vision

Image generation (Grok)

Generate gambar dengan Grok lewat POST /v1/responses + tool image_generation. Model: grok-4.5.

bash · generate image

python · generate image

• Endpoint: POST https://inferenesia.cloud/v1/responses

• Response berisi output[] tipe image_generation_call / output_image (base64)

• Opsional: lampirkan referensi image lewat input_image (data URL)

Anthropic Messages (Claude wire)

POST /anthropic/v1/messages

Format Anthropic asli (top-level system, max_tokens, streaming message_start/content_block_delta/message_stop). Cocok untuk SDK @anthropic-ai/sdk atau LangChain ChatAnthropic. Auth x-api-key (Bearer juga diterima). Model wajib enxapi/claude-*.

bash · non-stream

curl https://inferenesia.cloud/anthropic/v1/messages \
  -H 'x-api-key: ' \
  -H 'anthropic-version: 2023-06-01' \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "enxapi/claude-opus-5",
    "max_tokens": 128,
    "system": "Kamu asisten ringkas.",
    "messages": [{"role":"user","content":"Hi"}]
  }'

python · Anthropic SDK

from anthropic import Anthropic
client = Anthropic(api_key="", base_url="https://inferenesia.cloud/anthropic")
msg = client.messages.create(
    model="enxapi/claude-opus-5",
    max_tokens=256,
    system="Kamu ringkas.",
    messages=[{"role": "user", "content": "Hi"}],
)
print(msg.content[0].text)

• Streaming: kirim stream: true — respon SSE Anthropic format (bukan OpenAI delta).

• Model non-Claude → HTTP 400 anthropic messages endpoint is enxapi-only.

Fitur proxy (otomatis)

Token rotation

Round-robin / sticky pool di backend

RTK Token Saver

Compress tool output — hemat input tokens

Filters

Atur di menu Filters (per API key)

Streaming SSE

stream: true OpenAI-compatible

Admin API

Butuh JWT session dari login dashboard (bukan API key temp-)

POST /auth/loginLogin → JWT

GET /admin/accountsList provider accounts

POST /admin/accounts/importBulk import farm

GET/POST /admin/keysManage API keys

GET /admin/statusService health

Cara pakai

• Base URL: https://inferenesia.cloud/v1

• Auth: header Authorization: Bearer <api_key>

• Pilih tool di bawah, salin snippet, lalu tempel ke pengaturan client.

Generator integrasi

API key (sesi login)

Model

Base URL

https://inferenesia.cloud/v1

Snippet digenerate di browser. Ganti key/model sesuai kebutuhan.

Checklist

1. Buat API key di dashboard (prefix temp-)

2. Set base URL: https://inferenesia.cloud/v1

3. Pilih model id dari daftar di tab API

4. Untuk model CodeBuddy CN, aktifkan streaming

5. Model free default: grok-4.5

Session · · agent tools

Session logs

Inferenesia Cloud

Request Details

Create API key

Import accounts

API key created

Create user

Edit user

Import proxies