Integration

Venice.ai integration

Venice.ai is a private, OpenAI-compatible inference API. Generate chat completions and Responses-style answers with open and frontier text models (with tools, structured output, reasoning controls, web search and Venice characters), create embeddings, evaluate structured yes/no, choice and score questions, generate, edit, upscale and remove the background of images, turn text into speech, transcribe audio, search the web, scrape a page to markdown, extract text from documents, browse the model catalogue and public characters, and read the account's balance, usage and rate limits. API requests are pay-per-use, deducted from the account's DIEM, credits or USD balance. Not exposed: video, music and voice-changer generation (their result call answers JSON while running and the file once done, which this connector cannot read), voice cloning, streaming, API key management, the wallet (x402) payment and crypto RPC routes, and the OpenAI-compatible image alias. Authentication is a Venice API key sent as Authorization: Bearer. An inference-only key covers the model tools and get_rate_limits; Venice limits inference keys to the inference endpoints, so the balance and usage reads may need an admin key.

Connect Venice.ai How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Venice.ai.Built from the 26 Venice.ai tools below, on the connection you add once.

What you can do with Venice.ai

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.

Each badge shows the default agent permission of an endpoint (allowed: an agent calls it without asking; ask first: you are asked before an agent calls it, when the agent or chat asks permission for sensitive actions; never: no agent may call it), which every user can change endpoint by endpoint in their credential settings.

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

Image, 4 operations

edit_imageEdit imagePOSTAgents: ask firstEdits an image from text directions (for example remove the tree) with a Venice edit model and returns the edited image as a file. image is a file or an http(s) URL. Billed per edit.
edit_image_layersEdit image layersPOSTAgents: ask firstEdits a base image with extra layer or mask images from text directions and returns the edited image as a file. images is a list of URLs or base64 strings, base image first. Billed per edit.
generate_imageGenerate imagePOSTAgents: ask firstGenerates 1 to 4 images from a prompt with a Venice image model and returns them as files in images, with the request echo and timing. Billed per image; see list_models (type image) for prices.
upscale_imageUpscale imagePOSTAgents: ask firstUpscales an image file 2x or 4x (scale) and returns the larger image as a PNG file. Billed per call; prices per scale are in list_models (type upscale).

Character, 2 operations

get_characterGet characterGETAgents: ask firstPreview. Returns one public Venice character by its slug: name, description, model, tags, stats and the slug to pass as venice_parameters.character_slug in create_chat_completion.
list_charactersList charactersGETAgents: ask firstPreview. Lists public Venice characters (personas to chat with through venice_parameters.character_slug) with search, tag, category, model and sort filters, at most 100 per page.

Audio, 1 operation

transcribe_audioTranscribe audioPOSTAgents: ask firstTranscribes an audio file (WAV, FLAC, M4A, AAC, MP4, MP3, OGG, WEBM) into text in its own language, optionally with timestamps. Billed per audio duration.

Balance, 1 operation

get_billing_balanceGet billing balanceGETAgents: ask firstReturns the account's remaining balances (USD, DIEM for the current epoch, bundled and earned credits), the currency calls are billed in, whether it can make calls, and the DIEM epoch allocation.

Character review, 1 operation

list_character_reviewsList character reviewsGETAgents: ask firstPreview. Returns the public reviews of one character (rating, message, date, the reviewer's display name and avatar), with a rating summary and paging. Contains reviewers' public display names.

Chat completion, 1 operation

create_chat_completionCreate chat completionPOSTAgents: ask firstGenerates the next assistant message for a conversation (system, user, assistant, tool messages; images, audio, video or files for models that accept them). Returns choices[0].message, token usage and the cost. Billed per token.

Decision, 1 operation

evaluate_decisionsEvaluate decisionsPOSTAgents: ask firstBeta. Evaluates a state (text or JSON data) against typed questions and returns one structured answer per question id: yes/no probability (noul), a chosen option with probabilities (choice) or a rubric score.

Document, 1 operation

parse_documentParse documentPOSTAgents: ask firstExtracts the text of a document file (PDF, DOCX, PPTX, XLSX or plain text, at most 25 MB) and returns it with its token count. Processed in memory and not stored by Venice.

Embedding, 1 operation

create_embeddingsCreate embeddingsPOSTAgents: ask firstCreates embedding vectors for a string or an array of strings with a Venice embedding model, returning one vector per input in data[].embedding. Billed per input token.

Image background, 1 operation

remove_image_backgroundRemove image backgroundPOSTAgents: ask firstRemoves the background from an image file and returns the result as a PNG file with a transparent background.

Image style, 1 operation

list_image_stylesList image stylesGETAgents: ask firstLists the image style names generate_image accepts in style_preset (for example 3D Model, Anime, Cinematic).

Model, 1 operation

list_modelsList modelsGETAgents: ask firstLists Venice models of one type (text by default, or all) with id, name, context length, capabilities, constraints, privacy level and per-unit pricing in USD and DIEM.

Model mapping, 1 operation

list_model_compatibility_mappingsList model compatibility mappingsGETAgents: ask firstReturns the map of other providers' model names (such as gpt-4.1 or claude-opus-4-0) to the Venice model each one is served by; those names work as a model ID.

Model trait, 1 operation

list_model_traitsList model traitsGETAgents: ask firstReturns the model trait map for a model type (for example default, fastest, most_intelligent, default_vision), each trait naming the model ID it points to; a trait name works as a model ID.

Rate limit, 1 operation

get_rate_limitsGet rate limitsGETAgents: ask firstReturns the calling key's access status, API tier, balances, expiry, next epoch start, and the rate limits per model (requests per minute or day, tokens per minute).

Rate limit event, 1 operation

list_rate_limit_eventsList rate limit eventsGETAgents: ask firstReturns the last 50 times the account exceeded a rate limit: key id, model, tier, limit type (RPM, RPD, TPM, FAILED_REQUESTS, UNSUPPORTED_FEATURE_REQUESTS) and time.

Response, 1 operation

create_responseCreate responsePOSTAgents: ask firstBeta. Creates a model answer in OpenAI's Responses format from input (a string or messages) and optional instructions, returning typed output items (reasoning, message, function_call) and usage. Billed per token.

Speech, 1 operation

create_speechCreate speechPOSTAgents: ask firstTurns text (up to 4,096 characters) into speech with a Venice voice model and returns the audio as a file. Voices are model-specific. Billed per character.

Usage, 1 operation

get_usage_analyticsGet usage analyticsGETAgents: ask firstBeta. Returns aggregated usage over a lookback (up to 90 days) or a date range: totals by date, by model (with token breakdown) and by API key, in USD and DIEM. Data is cached for 10 minutes.

Usage entry, 1 operation

list_usage_historyList usage historyGETAgents: ask firstReturns billing ledger entries (amount, currency, SKU, unit price, tokens, API key id, timestamp) in ascending time order, a page at a time; pass nextCursor alone as cursor for the next page.

Web page, 1 operation

scrape_web_pageScrape web pagePOSTAgents: ask firstFetches a public web page and returns its content as markdown. Sites that block automated access (such as X/Twitter and Reddit) are refused at once.

Web search, 1 operation

search_webSearch webPOSTAgents: ask firstSearches the web and returns up to 20 results (title, URL, content snippet, date). Provider brave (default, zero data retention), google (anonymized through Venice) or exa.

Use Venice.ai from Claude Code, Cursor or Codex

Connect Venice.ai in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Venice.ai 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 Venice.ai

Create your workspace
Sign up free. Nothing to install.
Connect Venice.ai once
Add your Venice.ai 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.