LILT for AI agents
LILT provides AI-assisted translation and localization, letting teams translate text and files and manage trained language models tuned to their own content. Connect it once through Arc0, and your agent, or Claude, ChatGPT and Cursor, can use it through one MCP endpoint, limited to what each user approved.
What agents do in LILT.
Translate a piece of text
Translate a short text string into a target language for a document draft.
Check a translation job's status
Look up the status of a submitted file translation job to see if it is ready, a read-only check.
Approve before uploading a large file
Require approval before uploading a large file for translation, since it consumes account credits.
9 LILT actions, graded by risk.
Every LILT action is tagged read, write or destructive, so one policy covers the whole app and new actions inherit the right default.
read
6Look things up. Allowed by default.
- lilt.list_resourcesList resources
- lilt.get_credit_balance_informationGet credit balance information
- lilt.download_jobDownload job
- lilt.translate_textTranslate text
- lilt.check_job_statusCheck job status
- lilt.translate_files_with_verificationTranslate files with verification
write
3Create and change things. Allow, or ask the user first.
- lilt.create_trained_modelCreate trained model
- lilt.hello_worldHello world
- lilt.upload_fileUpload file
destructive
0Delete, cancel or archive. Ask first, or deny outright.
- No destructive actions.
LILT in three steps.
- 01Your users connect LILTThey sign in to LILT on Arc0 Connect, under your brand, and approve the access you ask for.
- 02You set the rulesReads run, and writes like “create trained model” can wait for the user to approve.
- 03Any agent can actYour agent calls LILT through the Arc0 SDK or MCP, and so can Claude, ChatGPT and Cursor. Every call lands on the audit log.
await arc0.policies.set('lilt', { read: 'allow', write: 'ask', // create_trained_model destructive: 'deny', }) # Claude Code: the same connection, one URL $ claude mcp add --transport http arc0 \ https://mcp.arc0.ai/u/u_8f2
How LILT connects.
LILT runs its own MCP server behind OAuth. Arc0 registers the client, users approve access on Arc0 Connect, and your agent reaches LILT through the same endpoint, policies and audit log as every other app.
The same LILT connection serves your agent over MCP and your own backend over REST and the proxy, so a user connects once. How Arc0 handles credentials →
- AUTH
- OAuth 2.0 (MCP)
- CREDENTIALS
- Per-tenant encrypted vault
- MODEL SEES
- Results only, never credentials
- AUDIT LOG
- Every call, on every plan
Use LILT from any agent.
ClaudeChatGPTCursor
Codex
VS CodeOpenAI Agents SDKClaude Agent SDK
Vercel AI SDK
Mastra
LangGraph
LILT and Arc0, answered.
Can I use LILT with Claude, ChatGPT or Cursor?
Yes. Connect LILT to Arc0 once, then add your Arc0 MCP URL to Claude, ChatGPT, Cursor, Claude Code or any other remote-MCP client. Each assistant only gets the LILT actions you allow.
How do users connect LILT?
LILT runs its own MCP server behind OAuth. Arc0 registers the client, users approve access on Arc0 Connect, and your agent reaches LILT through the same endpoint, policies and audit log as every other app.
Which LILT actions can my agent take?
9 in total: 6 read, 3 write and 0 destructive, such as “create trained model”. Your policies decide which of them each agent may call.
Can I make my agent read-only in LILT?
Yes. Allow read actions and deny writes in the LILT policy. Your agent can still look things up, and any write it attempts is blocked and logged.
Can my own backend call LILT too?
Yes. The same LILT connection is available over REST and through the Arc0 proxy, so your product and your agent share one connection per user.
Plug LILT into your agent.
Your users connect LILT once, under your brand. Your agent gets 9 actions behind your policies, with every call on the record.
Free to build · MCP + REST · Audit log on every plan