PureBridge API Docs

Request sample

POST /v1/chat/completions

Chat completions support stream: true for interactive responses. Use stream: false when you want a simple header-verification request.

Signed-in users see model IDs available to their current API plan. Public docs use the MODEL_ID placeholder; use the exact values returned by GET /v1/models.

curl https://api.purebridge.uk/v1/chat/completions \
  -H "Authorization: Bearer sk-xxxxxxxxxxxxxxxxxxxxxxxx" \
  -H "Content-Type: application/json" \
  -d '{
  "model": "YOUR_MODEL_ID",
  "messages": [
    {
      "role": "system",
      "content": "You are a concise assistant."
    },
    {
      "role": "user",
      "content": "Say hello from PureBridge."
    }
  ],
  "stream": true
}'

Response sample

Standard SSE streaming payload

data: {"id":"chatcmpl_123","object":"chat.completion.chunk","model":"YOUR_MODEL_ID","choices":[{"index":0,"delta":{"role":"assistant"},"finish_reason":null}]}

data: {"id":"chatcmpl_123","object":"chat.completion.chunk","model":"YOUR_MODEL_ID","choices":[{"index":0,"delta":{"content":"Hello from PureBridge."},"finish_reason":null}]}

data: {"id":"chatcmpl_123","object":"chat.completion.chunk","model":"YOUR_MODEL_ID","choices":[{"index":0,"delta":{},"finish_reason":"stop"}]}

data: [DONE]

Verifiable request

Check the official request id for OpenAI-routed calls

When the selected model routes to the official OpenAI upstream, PureBridge passes through the x-request-id returned by OpenAI. You can also send X-Client-Request-Id so your own trace id is available in upstream troubleshooting.

• x-request-id is the request ID generated by the OpenAI API for the upstream request. Record it for PureBridge-side troubleshooting and reconciliation.
• X-Client-Request-Id is generated by you and sent with the request, making it easier to connect local logs, PureBridge usage records, and upstream support traces.
• If the model ID routes to a non-OpenAI upstream, use that upstream's response headers and error format instead.
• The body sha256 in the sample is the hash of the bytes you sent locally. If gateway-level body-sha256 echoing is enabled later, it can be used to compare the upstream body as well.

verify.sh

Response-header verification sample

export PB_API_KEY="sk-xxxxxxxxxxxxxxxxxxxxxxxx"
CLIENT_REQUEST_ID="pb-verify-$(date +%s)"
BODY='{"model":"YOUR_MODEL_ID","messages":[{"role":"user","content":"ping"}],"stream":false}'

printf "%s" "$BODY" | shasum -a 256

curl -sS -D - -o /tmp/purebridge-verify.json https://api.purebridge.uk/v1/chat/completions \
  -H "Authorization: Bearer $PB_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Client-Request-Id: $CLIENT_REQUEST_ID" \
  --data-binary "$BODY" \
  | grep -Ei '^(x-request-id|openai-|x-ratelimit|date):'

Enabled model IDs

PureBridge model IDs available in this environment

Signed-in users see model IDs available to their current API plan. Public docs use the MODEL_ID placeholder; use the exact values returned by GET /v1/models.

Public docs do not display the platform-wide model directory. Sign in to the API Key workspace to see your plan models, then call GET /v1/models with your key.

Models payload

GET /v1/models response shape

{
  "object": "list",
  "data": []
}

Scenario focus

What this path is for

• OpenAI-compatible API usage: configure PureBridge in SDKs, CLI tools, IDE agents, or workflow tools.
• Developer productivity: Codex, opencode, Aider, Continue, Cline, Roo Code, and similar tools for code edits, explanations, tests, and automation.
• Personal automation: daily reports, document processing, message summaries, content generation, personal bots, scheduled scripts, and low-code workflows.
• Budget control: split keys by tool, project, or workflow, then track requests, tokens, cost, and unusual spend from usage records.

Next step

What to do after reading

Register an account, create an API KEY in App Workspace, then call the API domain.