LLM
A language model for the Agent: an EsaSim model, a provider with your key, or your own server.
How it works
Choose where the model runs. EsaSim models are billed per token from your credits. With a provider (Anthropic, OpenAI, Google, Mistral, xAI, DeepSeek) paste your API key once: it is stored encrypted in the workspace and only sent to the worker running the block. Custom: any OpenAI-compatible server, such as Ollama or vLLM on your own worker. Connect the output to an Agent's LLM input.
Outputs
| Output | Gives | What it is |
|---|---|---|
Modelllm | Model |
Settings
| Setting | Default | What it does |
|---|---|---|
Where it runsprovider |
esasim | One of: esasim, anthropic, openai, google, mistral, xai, deepseek, custom |
Modelmodel |
empty | |
API keyapi_key |
pasted in the editor, kept encrypted | |
Server address (custom)base_url |
empty | For example http://localhost:11434/v1 |
Temperaturetemperature |
0.2 | From 0 to 2 |
Max tokens per answermax_tokens |
2048 | From 16 to 200000 |
Where it runs
Its settings are used by the blocks connected to it; see Blueprints.
More in AI
AgentReasons over your data in a loop with the LLM you connect, using skills you allow.Check labelsFinds images whose label is likely wrong, from the labels of the images that look most like them, and images unlike any other.Embed imagesTurns every image into a vector with a vision model (CLIP or DINOv2) on the CPU, an NVIDIA GPU or Tenstorrent cards; images per second live; optional zero-shot labels.Text to imageTurns a prompt into any number of images with a diffusion model downloaded from Hugging Face onto the worker, from small (Segmind Vega, 0.7B) to large (FLUX.1 schnell, 12B): one image per seed, on an NVIDIA GPU, a Tenstorrent card or the CPU, each shown live as it is made.Text to videoTurns prompts into short videos (up to 10 seconds) with an open text-to-video model downloaded from Hugging Face onto the worker, from small (Wan 2.1, 1.3B) to the best open ones (Wan 2.2 A14B, and LTX-2.3 and LTX-2.5, which also make the sound): several variants of each prompt, one per seed, on the worker's NVIDIA GPU, each streamed live, whole, as soon as it is made.