Providers Overview

Providers are the LLM inference backends that meka uses to process your instructions. meka ships with four built-in backends, each selectable as a profile type:

Backend	Auth	API	Notes
`openai-api`	API key	Chat Completions	Works with OpenAI and any compatible endpoint (Ollama, vLLM, OpenRouter, …)
`openai-codex`	OAuth login	OpenAI Responses	Uses a ChatGPT subscription; talks to `chatgpt.com/backend-api/codex` like the Codex CLI
`claude-api`	API key	Claude Messages	Direct Claude API, billed per-token
`claude-oauth`	OAuth login	Claude Messages	Uses a Claude Code subscription; replicates Claude Code’s request shape and attestation

Configuring a Provider

Providers are configured as named profiles. The easiest way is meka provider add, which writes the profile to the config file and stores the secret (API key or OAuth token) in the database:

$ meka provider add work --type claude-oauth --model claude-opus-4-6

This produces a [providers.work] entry in ~/.config/meka/config.toml:

default_provider = "work"

[providers.work]
type  = "claude-oauth"
model = "claude-opus-4-6"

Selecting a Provider

meka uses the profile named by --provider <name> (per run), else default_provider, else the sole profile. Switch the default with meka provider use <name>:

meka --provider work     # this run only
meka provider use work   # persist as default_provider

There is no environment-variable override for provider selection.

OpenAI-Compatible APIs

The openai-api backend works with any API that implements the OpenAI Chat Completions format. This includes:

OpenAI (default endpoint)
Ollama (http://localhost:11434/v1)
OpenRouter (https://openrouter.ai/api/v1)
vLLM, LiteLLM, and other OpenAI-compatible servers

Set the profile’s base_url (or the --base-url flag for one run) to point at the alternative endpoint.

claude-api vs claude-oauth

Both talk to Claude’s /v1/messages endpoint, but the auth and request shape differ:

claude-api is the straightforward path: an x-api-key header, a plain system prompt, no extra headers. Choose this when you have a Claude API key.
claude-oauth replicates the Claude Code CLI exactly: OAuth tokens, fingerprint-encoded version header, xxHash64 attestation over the request body, injected billing system block. Choose this when you want to use a Claude Code subscription. Any deviation from the expected shape causes requests to be rejected, so avoid proxies that rewrite headers or reformat the body.

openai-api vs openai-codex

The two OpenAI-flavoured providers hit different endpoints with different protocols:

openai-api posts to /chat/completions on api.openai.com (or any compatible endpoint), authenticating with an API key. This is the right choice when you have an OpenAI billing account or are pointing at a self-hosted OpenAI-compatible server.
openai-codex posts to chatgpt.com/backend-api/codex/responses using the OpenAI Responses API (a different protocol: different request body shape, different streaming events). Authentication is OAuth against auth.openai.com, mirroring the first-party Codex CLI. Choose this to use a ChatGPT Plus / Pro / Team / Business subscription instead of a per-token API key.

Streaming vs Non-Streaming

By default, meka uses streaming mode: tokens appear in the terminal as they are generated. Use --no-stream to wait for the complete response before displaying it.

Streaming is recommended for interactive use. Non-streaming may be useful for scripting or when the provider does not support SSE.

Keyboard shortcuts