How do I use VIDEO AI ME with Codex?
Last updated:
To connect Codex to VIDEO AI ME, add an mcp_servers.videoaime block to ~/.codex/config.toml with url https://videoai.me/api/v1/mcp and an Authorization header carrying your Bearer API key. Create the key at videoai.me/developers on a Pro or Premium plan. Codex can then generate video ads and publish them to social platforms.
How do I connect Codex to VIDEO AI ME?
1.Create an API key
Log in at videoai.me and open /developers (Developers in the sidebar). Create a key, optionally with a budget cap in dollars so an agent can never overspend, and copy the full va_live_... value immediately: it is shown exactly once. API keys require an active Pro ($99/month) or Premium ($199/month) subscription.
2.Add the server to config.toml
Codex reads MCP servers from ~/.codex/config.toml. The VIDEO AI ME server is remote streamable HTTP, so declare the url and the Authorization header:
~/.codex/config.toml
[mcp_servers.videoaime] url = "https://videoai.me/api/v1/mcp" http_headers = { "Authorization" = "Bearer va_live_YOUR_KEY" }3.Confirm the tools and run one
Start a new Codex session and list the available tools; the videoaime skills should appear. Begin with a free discovery skill, for example list_templates, then ask for a maker run. Codex receives the credit estimate first and asks you to confirm before the run starts.
What can Codex do once connected?
Every skill in the VIDEO AI ME registry: maker skills that generate UGC-style ads, product ads, explainers, and talking-head clips, publishing skills that post and schedule finished videos to your connected social accounts, free discovery skills for templates, actors, voices, and brand memory, and check_run for polling long runs. The full catalog with credit costs and typical durations is on the skill center page, and the REST API equivalent of every tool is described at https://videoai.me/api/v1/openapi.json.
What does access cost?
The API and MCP server are included with the Pro ($99/month) and Premium ($199/month) plans; there is no separate API price. Maker skills spend video credits from the plan's monthly allowance at 1 credit = 1 cent, always quoting the estimate first, and each API key can carry its own budget cap in dollars. Plans and allowances are on the pricing page.
Codex setup questions
Direct answers about connecting Codex to VIDEO AI ME.
Does Codex support remote MCP servers like the VIDEO AI ME one?
Yes. Codex supports streamable HTTP MCP servers declared in ~/.codex/config.toml with a url field, and custom headers via http_headers, which is exactly what the VIDEO AI ME server needs: the hosted endpoint plus an Authorization Bearer header. No local package or wrapper process is required.
How should I handle the API key in Codex configuration?
Keep ~/.codex/config.toml out of version control and readable only by your user, since it contains the key. If the key is ever exposed, revoke it at videoai.me/developers and create a replacement; the old key stops working immediately, and only its prefix remains visible in the key list.
Can I run VIDEO AI ME skills from Codex in CI or automation?
Yes, with two cautions. Maker skills answer with needs_confirmation before spending credits, so an unattended pipeline must quote and confirm deliberately, not blindly. And give automation its own key with a budget cap in dollars, so a runaway job is stopped by the cap instead of draining the monthly allowance.
What error do I get when a key hits its budget cap in Codex?
A key_budget_exceeded error with a 402 status, stating the estimated cost of the refused run and the credits left under the cap. Free skills keep working. To continue, raise the cap on that key, or create a new key; both take a few seconds in the dashboard.
How does Codex know what each VIDEO AI ME skill costs?
Costs are part of the skill metadata itself. Every skill description carries its credit cost range (1 credit = 1 cent), the same numbers published at videoai.me/skills, and maker skills return an exact estimate for your specific input before running. The OpenAPI spec at https://videoai.me/api/v1/openapi.json carries the same schemas for the REST API.