Overview
Ollamancer is a fully local, terminal-first AI agent for Ollama. It is built from scratch and focuses on honesty, small-model reliability, and privacy. The project states that no cloud, API keys, or data leaving the machine are required; users point it at a project folder and talk to it. It reasons and acts with 35 native tools, MCP servers, and the shell. It includes local RAG, vision, dual-model planning, skills, and a terminal UX. It also provides deterministic honesty controls, small-model reliability engineering, and a --private ephemeral mode that writes nothing to disk. The project is v3.1, developed on macOS (Apple Silicon, 24 GB), requires Python 3.12+, and is MIT licensed.
Pricing
Free
Detailed plans have not been confirmed in our catalog. Check the official website for current limits and billing terms.
Visit WebsitePrices and limits may change. Confirm the currency, billing period, seat minimum and usage caps on the official website.
Use Cases
- Users use Ollamancer to fix failing tests in a project and verify by running them.
- Users use Ollamancer to find where retry logic is handled using local RAG.
- Users use Ollamancer to perform a security review of their changes.
- Users use Ollamancer to research a topic and write a cited report.
- Users use Ollamancer to run headless or batch prompts for cron and scripts.
- Users use Ollamancer to load reusable SKILL.md workflows for tasks such as test-and-fix or security-review.
- Users use Ollamancer to describe screenshots or read charts via an installed multimodal model.
- Users use Ollamancer to keep a persistent Python REPL for real data work.
Features
Fully local terminal AI agent
Works with Ollama models
Includes 34 agent tools
Supports Model Context Protocol
Local retrieval-augmented generation
Deterministic hallucination checks
Runs without cloud or API keys
Reviews
No reviews yet. Be the first to share your experience!