Providers

Provider-agnostic: same tools, same sessions across Anthropic, OpenAI, and Gemini.

Meerkat is provider-agnostic. Your tools, sessions, hooks, and configuration work identically across Anthropic, OpenAI, and Gemini. Switch providers by changing the model name — no code changes. Set an API key for at least one.

Provider setup

Anthropic

export ANTHROPIC_API_KEY="sk-ant-..."

Model	Context	Max output	Best for
claude-opus-4-6	200K / 1M (beta)	128K	Complex reasoning, highest quality
claude-sonnet-4-5	200K / 1M (beta)	64K	Balanced performance and cost
claude-opus-4-5	200K	64K	Legacy Opus (still supported)
claude-haiku-4-5	200K	64K	Fast, simple tasks

.rkat/config.toml

[agent]
model = "claude-opus-4-6"
max_tokens_per_turn = 16384

OpenAI

export OPENAI_API_KEY="sk-..."

Model	Context	Best for
gpt-5.2	1M	General purpose, advanced reasoning
gpt-5.2-pro	1M	Highest quality reasoning
gpt-5.1-codex-max	1M	Code generation, agentic coding

.rkat/config.toml

[agent]
model = "gpt-5.2"
max_tokens_per_turn = 8192

Gemini

export GOOGLE_API_KEY="AIza..."

Model	Context	Best for
gemini-3-pro-preview	1M	Advanced reasoning, complex tasks
gemini-3-flash-preview	1M	Fast, balanced performance

.rkat/config.toml

[agent]
model = "gemini-3-pro-preview"
max_tokens_per_turn = 8192

Environment variables

Variable	Fallback	Provider
RKAT_ANTHROPIC_API_KEY	ANTHROPIC_API_KEY	Anthropic Claude
RKAT_OPENAI_API_KEY	OPENAI_API_KEY	OpenAI GPT
RKAT_GEMINI_API_KEY	GEMINI_API_KEY, GOOGLE_API_KEY	Google Gemini

The RKAT_* variants take precedence over provider-native names, so you can run Meerkat with dedicated keys separate from other tools.

SDK feature flags

When using Meerkat as a Rust library, enable only the providers you need:

Feature	Description	Default
anthropic	Anthropic Claude support	Yes
openai	OpenAI GPT support	No
gemini	Google Gemini support	No
all-providers	All LLM providers	No

Anthropic only (smallest binary)

meerkat = { version = "0.1", features = ["anthropic", "jsonl-store"] }

Provider parameters

Provider-specific options can be passed via the --param CLI flag or provider_params in the SDK:

Anthropic

Parameter	Description
thinking_budget	Token budget for extended thinking (integer)
top_k	Top-k sampling parameter (integer)

rkat run --model claude-sonnet-4-5 --param thinking_budget=10000 "Solve this problem"

OpenAI

Parameter	Values	Description
reasoning_effort	low, medium, high	Reasoning effort for o-series models
seed	Integer	Seed for deterministic outputs

rkat run --model gpt-5.2 --param reasoning_effort=high "Prove this theorem"

Gemini

Parameter	Description
thinking_budget	Token budget for extended thinking (integer)
top_k	Top-k sampling parameter (integer)

rkat run --model gemini-3-flash-preview --param thinking_budget=5000 "Analyze this data"

Auto-detection

The provider is automatically inferred from the model name:

• claude-* models use Anthropic
• gpt-*, o1-*, o3-*, chatgpt-* models use OpenAI
• gemini-* models use Gemini

You can override this with --provider on the CLI or provider in API requests.