AI and agents · by Ollama Inc
Ollama
Chat, search by meaning or speak text aloud, all offline.
text, image → text, data · text → embedding · text, data, image → data
- Way in
- Official APINo sign-up, no key
- To start
- Freefree; cloud models need an ollama.com plan
- Runs
- On your Mac
- Checked
- Nov 3, 2026Prices as of that day
What it can do 6
Chat with a local LLM (tools, thinking, structured JSON output)RAM-bound: on 16 GB keep model+KV under ~10 GB (e.g. qwen3.5:9b 6.6 GB, gemma4:12b-it-qat 7.2 GB)text, image → text, datafree (local)
Embed text for search/RAGtext → embeddingfree (local)
Answer typed choice / yes-no / score questions about a text or image state in one forward pass (decision models)Ollama >= 0.35 (/v1/systemone); local models nimble 9B, tev1 4B/0.8B, clef-flash 9B, d3, kev, ...text, data, image → datafree (local)
Speak text aloud (Phonon TTS, streaming WAV)macOS only for now; 24 kHz monotext → speechfree (local)
Generate an image from a prompt (experimental, macOS)x/flux2-klein 4B fp4 = 5.7 GB fits; x/z-image-turbo = 13 GB does not fit in 16 GB comfortablytext → imagefree (local)
Route requests to Ollama cloud models (:cloud tags) through the same local APItext → textPay-as-you-go $0 + per-token; Pro $20/mo (incl. $60 credits), Max $100/mo (incl. $300 credits); per-million-token rates on the pricing page
Does the same job
Prices and limits as of Nov 3, 2026, from Ollama Inc’s own pages. They change; check before you rely on them.Not affiliated with Ollama Inc. Not advice.Something wrong?