news Codistry

Codistry - The coding agent that understands

Codistry is a VS Code coding assistant designed to understand an entire codebase, not just isolated files. It maps symbols, relationships, and change patterns across the repository so that edits propagate through related callers, tests, and dependencies. The product emphasizes saving tokens by explaining the codebase once and then sending only the relevant context for each task. It cites the code it reads and keeps its index updated as commits are made. A key feature is its context engine, which builds a knowledge graph of the repo and identifies related files for a given problem. In one example, it traces an OAuth session issue to the files involved in key rotation and session behavior, then packages only those files for the model. Codistry also adapts to user habits by learning from commit history, matching the user’s conventions, expertise, and preferred level of explanation. The tool supports multiple model options, including hosted, local, and air-gapped setups, and is not tied to one provider. Users can switch models mid-session without reconfiguration. It integrates with existing workflows through parallel sessions, subagents, MCP integrations, and sandboxed skills. Getting started is simple: install the VS Code extension, connect a model, and give it a real coding task. A free plan is available, and no credit card is required.

Build this article’s timeline

Extract dated events, key actors, and open questions using the same timeline analysis available in Signal.

Log in to build

Timeline generation is available to signed-in community members.

Five Ws 🤖 3 AI prefills · 2 still unanswered
Who
Claude Haiku, Claude Opus, Ollama, and Codistry
🤖 AI prefill
What
Codistry is a VS Code coding assistant designed to understand an entire codebase, not just isolated files.
🤖 AI prefill
When When did this happen, or what is the timeframe?
Where Where did this occur?
Why
Centers on context engine, knowledge graph, and codebase mapping
🤖 AI prefill

Login to answer and earn XP for each gap you fill.

Codebase Context Engine codebase mapping, knowledge graph, symbols, relationships, dependencies Adaptive Coding Assistance coding agent, learns your habits, expertise, commit history, conventions Model Flexibility Options hosted models, local models, API keys, Claude, Ollama Developer Workflow Features VS Code, parallel sessions, subagents, MCP integrations, sandboxed skills Onboarding And Pricing install, pricing, free plan, no credit card, quick start
context engine98% knowledge graph95% codebase mapping93% tokens saved90% VS Code89% expertise mined from git history88% parallel sessions86% subagents85% MCP integrations84% sandboxed skills83%
6
Manipulation Score: Low How much this article uses emotional framing vs. straight reporting
Loaded Language 425%
Absolutist 815%
Threat/Panic 225%
Us vs Them 315%
Engagement Bait 1020%
Manipulation score look right?

Factual Claims (10)

medium technology
Codistry maps your entire codebase so every change carries through to every caller, test, and dependency.
medium technology
Codistry learns how you work and adapts to your codebase and habits with every session.
medium technology
Codistry explains your codebase once and then gives the model only the code needed for each task.
low technology
Codistry can be used in VS Code.
low economic
A free plan is available and no credit card is required.
medium technology
Codistry can switch between models mid-session without reconfiguration.
medium technology
Codistry supports hosted models, local models, and air-gapped offline use.
medium technology
The context engine maps a repo into a knowledge graph of symbols, relationships, and change history.
medium technology
Parallel sessions can run side by side without canceling each other.
medium technology
Sandboxed skills run inside isolated Docker containers with no host access.

Comments

Login to view and post comments.

🏠Portal 📰Links ❓Q&A 📅Events 💼Jobs