ctx
Open-source local CLI that indexes coding-agent transcripts from Claude Code, Codex, Cursor, OpenCode, Gemini CLI, Copilot CLI, and related harnesses for cited history search.
Do not bounce yet
Read the fit check, compare one alternative, then decide whether the vendor page is still your best next click.
Quick Verdict
Fast fit check before you leave the page
Make the fit call first. Vendor pages are good at selling, but they rarely tell you where the product is a bad match.
- Developers who run multiple coding-agent sessions per week and lose useful context between them
- Claude Code, Codex, Cursor, OpenCode, Gemini CLI, and Copilot CLI users who want local transcript search
- Teams building repeatable agent workflows where prior decisions and failed commands should be cited
- Local transcript text can contain sensitive paths, prompts, tool output, or secret-shaped strings, so copied search output still needs review.
- Provider import fidelity depends on each agent harness's local history format, and those formats can change without notice.
- History search can recover old context, but it cannot prove that the current repository still matches the previous session's assumptions.
Compare Next
Take one more internal step before the vendor pitch
This is where visitors usually jump out too early. Read one deeper take or open one alternative so the next click is informed instead of impulsive.
Alternative profile
Claude-Mem
Open-source persistent memory layer for Claude Code and other coding agents that captures session observations, compresses them, and injects relevant context back into future work.
Alternative profile
agentmemory
Persistent cross-session memory layer for coding agents, with MCP, hooks, and shared recall across Claude Code, Codex CLI, Gemini CLI, OpenClaw, and more.
Alternative profile
Recall
Fully local Claude Code memory plugin that logs sessions and writes compact resume context without sending project data to an external model.
ctx is worth tracking because serious vibe coding creates a hidden memory problem. The useful details are often in last week's agent transcript: why a migration failed, which command was already tried, what file path mattered, or which approach was rejected. ctx indexes those local coding-agent histories into SQLite and gives current agents a cited way to search prior sessions before they repeat the same work.
ctx is a Rust CLI for searching the coding-agent history already stored on a developer's machine. Instead of asking every new Claude Code, Codex, Cursor, OpenCode, Gemini CLI, Pi, Copilot CLI, or Factory Droid session to rediscover old decisions, failed commands, file paths, and constraints from scratch, ctx imports local transcript records into a private SQLite index and returns ranked matches with session and event IDs. It also ships an agent-history skill, read-only MCP server, CLI docs, SQL inspection, and SDK guidance, making it a practical context-recovery layer for vibe-coding teams that want agents to cite prior work before repeating it.
Why Choose ctx?
Choose ctx when your team uses Claude Code, Codex, Cursor, OpenCode, Gemini CLI, or similar agents often enough that old transcript context matters.
The strongest fit is local context recovery: decisions, failed attempts, touched files, commands, and constraints that are too detailed for a short memory note.
Cited session and event IDs make retrieval more inspectable than generic memory summaries because the current agent can open the underlying transcript slice.
Be cautious with privacy. ctx is local-first, but its index and copied output can preserve sensitive paths, prompts, tool output, or secret-shaped strings.
Key Features
Imports persisted local coding-agent histories into a private SQLite index instead of relying on fragile manual summaries.
Searches prior sessions with natural-language terms, file filters, multiple terms, and cited session or event IDs that a current agent can inspect before acting.
Supports local history sources for Claude Code, Codex, Cursor, Pi, OpenCode, Antigravity/Gemini CLI, Factory AI Droid, Copilot CLI, and several adjacent harnesses as provider adapters mature.
Ships an installable agent-history search skill for Codex, Claude Code, Cursor, and raw Agent Skills setups.
Provides a read-only MCP server through `ctx mcp serve` so compatible agents can query local history without copying the whole transcript corpus into context.
Includes local docs, CLI reference, read-only SQL inspection, SDK guidance, and signed installer-managed upgrades for the official binary path.
Pros & Cons
- It targets a real vibe-coding bottleneck: agents repeat old investigations because they cannot recover the conversation where a decision or failed command happened.
- Cited session and event IDs are more reviewable than opaque memory summaries because the agent or human can inspect the underlying transcript window.
- Local-first indexing and no default model/API calls make it easier to evaluate for privacy-sensitive developer workflows than hosted memory products.
- The provider matrix is directly relevant to current coding-agent stacks instead of being a generic notes app with an AI label.
- Local transcript text can contain sensitive paths, prompts, tool output, or secret-shaped strings, so copied search output still needs review.
- Provider import fidelity depends on each agent harness's local history format, and those formats can change without notice.
- History search can recover old context, but it cannot prove that the current repository still matches the previous session's assumptions.
- The project is young, with strong early signal but limited long-term evidence across large team installations.
Detailed Use Cases for ctx
Search old agent sessions for the failed migration, command output, or rejected fix before asking a new agent to continue the work.
Use file-aware search when a current task touches code that an earlier Claude Code, Codex, Cursor, or OpenCode session already investigated.
Install the bundled skill or read-only MCP server so compatible agents can query local history and cite session or event IDs in their answers.
SQLite storage, JSON output, docs, SQL inspection, and SDK guidance make ctx useful for structured retrieval without stuffing whole transcript folders into model context.
Who Should Use ctx?
Developers who run multiple coding-agent sessions per week and lose useful context between them
Claude Code, Codex, Cursor, OpenCode, Gemini CLI, and Copilot CLI users who want local transcript search
Teams building repeatable agent workflows where prior decisions and failed commands should be cited
Privacy-conscious builders who prefer local SQLite indexing over hosted agent-memory services
Perfect For
Letting a new Claude Code, Codex, Cursor, or OpenCode session find the prior debugging attempt before repeating the same failed path.
Searching previous agent sessions that touched a specific file before asking an agent to refactor it again.
Giving agents a cited local memory layer for decisions, constraints, rejected approaches, and command output from earlier work.
Auditing a personal or team workstation's local agent history with SQL and structured session metadata instead of raw transcript grep.
Technical Details
ctx Comparisons & Alternatives
ctx review
ctx vs Recall
ctx vs Claude-Mem
ctx vs Callimachus
local coding agent history search
Claude Code transcript search
Codex history search
agent memory SQLite CLI
Developers compare ctx with other vibe coding tools when they need a better workflow fit, not just a better landing page.
Direct Competitors
Recall
Callimachus
Claude-Mem
AgentMemory
grep or ripgrep over raw transcripts
Manual MEMORY.md notes
Similar Tools You Might Like
Rust-based coding-agent harness with a native TUI, multi-session workflows, agent memory, swarm coordination, MCP, browser tooling, and broad model-provider support.
Community Codex CLI fork that adds Auto Drive orchestration, background Auto Review, browser/CDP integration, theming, and multi-agent commands.
Open-source memory-first coding agent that turns disposable coding sessions into long-lived agents with persistent memory, skills, search, and multi-channel access.
Alternative Tools to Consider
Open-source persistent memory layer for Claude Code and other coding agents that captures session observations, compresses them, and injects relevant context back into future work.
Persistent cross-session memory layer for coding agents, with MCP, hooks, and shared recall across Claude Code, Codex CLI, Gemini CLI, OpenClaw, and more.
Fully local Claude Code memory plugin that logs sessions and writes compact resume context without sending project data to an external model.
Do one more comparison before you commit to ctx
Strong picks usually survive one more internal check. Read deeper, compare a neighbor, then leave for the vendor page if the fit still holds.