Quickstart

Get access

1. Request a rate card. Tell us the models and rough monthly volume. You get written per-model rates within one business day, AEST.
2. Receive your console login. Seedrise opens your company account on console.seedrise.ai. There is no self-serve signup.
3. Create a key. In the console, create keys per project or environment. Each key can be limited by budget, models, expiry and IP allow-list.

Endpoints and keys

Send your key as Authorization: Bearer sk-... on every request. Anthropic-style clients may send it as x-api-key instead; both work. The Responses API (/v1/responses) is not available yet.

Quickstart

A first request in curl. Swap the model for any ID on your rate card.

curl https://api.seedrise.ai/v1/chat/completions \
  -H "Authorization: Bearer sk-..." \
  -H "Content-Type: application/json" \
  -d '{"model":"gemini-3.5-flash","messages":[{"role":"user","content":"Hello"}]}'

# list the models your key can use
curl https://api.seedrise.ai/v1/models -H "Authorization: Bearer sk-..."

Claude Code

Full guide with team settings and troubleshooting: Claude Code on Seedrise.

Claude Code talks Anthropic Messages, so it points at the root URL. Set the environment once and start it as usual.

export ANTHROPIC_BASE_URL="https://api.seedrise.ai"
export ANTHROPIC_AUTH_TOKEN="sk-..."                     # your Seedrise key
export ANTHROPIC_API_KEY=""                             # clear any personal key
export ANTHROPIC_DEFAULT_OPUS_MODEL="claude-opus-5-5"
export ANTHROPIC_DEFAULT_SONNET_MODEL="claude-sonnet-5-5"
export ANTHROPIC_DEFAULT_HAIKU_MODEL="claude-haiku-5-5"
export CLAUDE_CODE_SUBAGENT_MODEL="claude-sonnet-5-5"
export CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1
claude

• Do not set any CLAUDE_CODE_USE_* cloud-provider flag. The client speaks plain Anthropic format to the gateway.
• If you were logged in with a personal Anthropic account, run /logout once so the cached login does not override the gateway.
• /status inside Claude Code should show the Seedrise gateway URL and your auth token.
• On start Claude Code calls GET /v1/models to fill the model picker. A 404 from count_tokens is expected; Claude Code falls back to its own estimate.
• If a beta header is rejected, set CLAUDE_CODE_DISABLE_EXPERIMENTAL_BETAS=1.
• Fable 5.1 is available on keys that have it enabled. Set ANTHROPIC_DEFAULT_FABLE_MODEL="claude-fable-5-1" on those keys only.

To lock a team to the gateway, ship this in the managed settings.json:

{
  "allowedProviders": ["customEndpoint"],
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.seedrise.ai",
    "ANTHROPIC_DEFAULT_SONNET_MODEL": "claude-sonnet-5-5"
  }
}

WorkBuddy

Full guide: WorkBuddy on Seedrise.

WorkBuddy uses OpenAI Chat Completions. Add Seedrise as a custom model.

Claude models are translated to Anthropic Messages on the gateway, so tool calls and streaming work without changes on your side.

DeepSeek Harness

Full guide: DeepSeek Harness on Seedrise.

In DeepSeek Harness, add a custom model service and choose the Anthropic Messages protocol.

Other tools

Anything that accepts a custom OpenAI-compatible endpoint works the same way: base URL, key, model ID.

Clients that only speak the Responses API (/v1/responses) cannot connect yet. Ask us if you depend on one.

OpenAI and Anthropic SDKs

OpenAI SDK, Python

from openai import OpenAI

client = OpenAI(base_url="https://api.seedrise.ai/v1", api_key="sk-...")

r = client.chat.completions.create(
    model="claude-sonnet-5-5",
    messages=[{"role": "user", "content": "Summarise this contract in three bullets."}],
)
print(r.choices[0].message.content)

OpenAI SDK, TypeScript, streaming

import OpenAI from "openai";

const client = new OpenAI({ baseURL: "https://api.seedrise.ai/v1", apiKey: process.env.SEEDRISE_API_KEY });

const r = await client.chat.completions.create({
  model: "deepseek-v3.2",
  messages: [{ role: "user", content: "Write a release note for v2.1." }],
  stream: true,
});
for await (const chunk of r) process.stdout.write(chunk.choices[0]?.delta?.content ?? "");

Anthropic SDK, Python

import anthropic

client = anthropic.Anthropic(base_url="https://api.seedrise.ai", api_key="sk-...")

msg = client.messages.create(
    model="claude-opus-5-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Explain data residency to a CFO."}],
)
print(msg.content[0].text)

Embeddings, image and video endpoints follow the same OpenAI request shape. Ask for the endpoint reference for your rate card.

Model IDs

Use the provider's own ID, no vendor prefix. The full list with prices, context windows and regions is in the model library. The IDs your key can call come back from GET /v1/models.

Regions and data

• Models marked AU can be pinned to Sydney or Melbourne on request; this is set per key, not per request.
• Prompts and outputs are never used to train a model. Seedrise does not retain request or response content; request metadata is kept for billing and audit.
• Zero-retention routing and a data processing agreement are available on request. Details on the data residency section.

Limits and errors

Default rate limits are set per key when your account is opened and raised on request. Long requests stream; keep client timeouts at 10 minutes or more for reasoning models.

Support

Email [email protected] with your company name and, for errors, the request ID from the response headers. Replies within one business day, AEST. Volume and Committed accounts have a named contact and a shared channel.