Integration

Arbitr integration

Arbitr (formerly Straker, whose Straker Verify product it continues) translates and verifies documents. Create a translation project from a file with an AI stage and optional human translation and edit stages, poll its status, read its verification findings and provenance record, download the translated deliverables one by one or as a ZIP, resume a project parked for lack of credits, and read the supported languages and the credit balance. Buying credits has no API and is done in the arbitr web app. The chat-app connection, usage export and short-text translation routes the docs announce are not yet served by the API host and are left out. Authentication is an API key from Settings, API Keys, sent in the X-API-Key header; abr_live_ keys do real (billed) work, abr_test_ keys only check requests and never translate. Each key carries scopes (verify:submit, verify:read, audit:read) chosen at creation. Rate limit: 100 requests per minute on a live key, 30 on a test key, 20 per second burst.

Connect Arbitr How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Arbitr.Built from the 14 Arbitr tools below, on the connection you add once.

What you can do with Arbitr

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Deliverable, 4 operations

download_deliverableDownload deliverableGETDownloads one deliverable file of a project and returns it as a stored file. Take the deliverable id from list_deliverables. Needs the verify:read scope.
download_deliverables_zipDownload deliverables zipGETDownloads every current deliverable of a project as one ZIP archive and returns it as a stored file. Call it once get_project reports completed. Needs the verify:read scope.
get_deliverableGet deliverableGETReturns one deliverable's metadata: file id, file type, name and locale. To get the file itself use download_deliverable. Needs the verify:read scope.
list_deliverablesList deliverablesGETLists the deliverable files of a project (id, file type, name, locale). Pass an id to download_deliverable for the file itself, or use download_deliverables_zip for all of them in one ZIP. Needs the verify:read scope.

Project, 3 operations

create_projectCreate projectPOSTCreates a translation project from one source file and starts it, charging credits (a test key charges nothing). Poll get_project with the returned id until it is done, then list_deliverables. Needs the verify:submit scope.
get_projectGet projectGETReturns one project's status, assessment (billable words, credit sufficiency, due date feasibility) and review (human stage). Poll it after create_project; check review.status before status: status can read completed when the human stage never ran.
list_projectsList projectsGETLists the key's projects, newest first, one page at a time: pass page (from 1) until page.has_more is false. Filter by status or by modified_after for incremental sync. Rows carry assessment and review as null: read them with get_project.

Resumption, 2 operations

resume_projectResume projectPOSTResumes a project parked at status awaiting_payment for lack of intelligence credits, spending them now. Top up in the arbitr web app first (no API does that). 402 if still short, 409 if the project is not parked.
resume_project_reviewResume project reviewPOSTRestarts a project's human stage when review.status is awaiting_payment, spending trust credits now. Top up in the arbitr web app first (no API does that). 402 if still short, 409 if not parked (also when review.status is skipped).

Balance, 1 operation

get_credit_balanceGet credit balanceGETReturns the organisation's remaining intelligence credits (AI stage) and trust credits (human stages), and usage in the current billing cycle. Any valid key may call it; a test key always reads 999,999.

Chain of custody, 1 operation

get_project_chain_of_custodyGet project chain of custodyGETReturns a project's provenance record: who created it and through which channel, the source files submitted, and every deliverable generation including superseded ones. Needs the audit:read scope. Carries no content hash.

Finding, 1 operation

list_project_findingsList project findingsGETLists a project's verification findings (segment flags and agent findings). Needs the audit:read scope. Pass page.after back as after until page.has_more is false; severity, category and status filter flags only and drop agent findings.

Language, 1 operation

list_languagesList languagesGETLists the languages arbitr supports, as lowercase BCP-47 codes (bcp47) and display names. Use these exact codes in create_project; other forms are refused. Any valid key may call it.

Me, 1 operation

get_api_key_infoGet api key infoGETReturns the organisation id, the mode (live or test) and the scopes of the API key in use. Any valid key may call it: use it to see which tools the key is allowed to call before trying them.

Use Arbitr from Claude Code, Cursor or Codex

Connect Arbitr in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Arbitr operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Arbitr

Create your workspace
Sign up free. Nothing to install.
Connect Arbitr once
Add your Arbitr connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.