avatar.llmer.com

Paid avatar renders for agent-native workflows.

Validate inputs, pay through x402, stream back a square MP4. No user accounts, no durable media storage, no public GPU surface.

Pricing /api/v1/pricing Capabilities /api/v1/capabilities Agent skill sheet /skills.md Hosted MCP helpers /mcp

local ascii field online

Tiers

Base

$0.25

Text + image only.

Up to 180 chars, 40 words, predicted speech under 8s.

Pro

$0.50

Optional reference audio.

Up to 450 chars, 100 words, predicted speech under 16s.

Extended

$1.00

Longer speech window with reference audio support.

Up to 900 chars, 200 words, predicted speech under 30s.

Flow

  1. Call POST /api/v1/validate with multipart form data.
  2. Pick the cheapest compatible tier or use the recommended one.
  3. Call /api/v1/render/{tier} with the same payload and an Idempotency-Key.
  4. Pay the x402 challenge, then receive video/mp4.

Agent Paths

Direct HTTP

Canonical paid surface for scanners and x402-native clients.

Hosted MCP

Read-only helper tools for pricing, capabilities, and validation.

Local MCP Bridge

Recommended for Codex, Claude, and OpenClaw. It handles payment and saves the returned MP4 locally.

No Retention

Requests are processed in memory or request-scoped temp space. Raw media and finished videos are not durably stored by the service.

The only persistent records are payment and refund metadata needed for reconciliation.

Quickstart

curl -X POST http://localhost:8080/api/v1/validate \
  -F 'text=hello world' \
  -F 'image=@./avatar.png'

uv run python -m avatar_llmer_mcp \
  --api-base-url http://localhost:8080