Prompt Deck vs Claude Code History Viewer, Agent Sessions and SpecStory
Checked against each tool's own site, docs and GitHub on 22 October 2026. Numbers in brackets link to the source.
Coding agents keep your conversations on disk, each in its own format. These four tools help you read them again. Claude Code History Viewer and Agent Sessions are free, MIT-licensed history browsers. SpecStory saves your chats as Markdown inside your projects. Prompt Deck reads the history too. It also keeps the prompts and memory files you reuse across agents.
Swipe the table sideways to see all four apps.
| Prompt Deck | Claude Code History Viewer | Agent Sessions | SpecStory | |
|---|---|---|---|---|
| What it is | A Mac app that puts history from 11 agents in one list, next to your prompts and memory files | A desktop app, or a self-hosted web page, for browsing and searching history from 31 assistants [1] | A Mac app to search, read and resume sessions from 16 agents, with a Codex and Claude usage meter [5] | Editor extensions and a command-line tool that save agent chats as Markdown in your project, with an optional cloud [9] |
| Runs on | macOS 14 or later, Apple silicon and Intel | macOS 10.13 or later, Windows, Linux [1] [4] | macOS 14 or later [8] | Inside Cursor and VS Code; the command-line tool installs on macOS, Linux and Windows [9] |
| Kind of app | Native (SwiftUI) | Tauri: Rust with a web view [1] | Native (Swift and SwiftUI) [6] | Editor extension, command-line tool and web cloud [9] |
| Agents it reads | 11: Claude Code, Codex CLI, Cursor, OpenCode, Antigravity CLI, GitHub Copilot in VS Code, Qwen Code, DeepSeek Harness, Continue, Pi, Aider | 31, including 10 of Prompt Deck's 11 (not DeepSeek Harness), plus Gemini CLI, Cline, Kiro, Goose and Zed [1] | 16; reads Copilot CLI but not Copilot in VS Code, Continue or Aider [5] | 14 plus Cursor Cloud Agents, including Copilot in VS Code and Gemini CLI [9] |
| Search across agents | Yes | Yes [1] | Yes [5] | Yes, from the command line; cloud search covers synced sessions only [10] [13] |
| Sub-agent conversations | Yes, opened inline under the call that started them | Yes [2] | Yes, nested under the parent session [6] | Not stated in its docs |
| Edit memory files | Yes: CLAUDE.md, AGENTS.md, GEMINI.md, QWEN.md, .cursorrules and Copilot instructions, global and per project | No memory-file editor; it edits Claude Code settings such as permissions, hooks and MCP [3] | No, read-only by design [7] | No editor; an opt-in feature writes Cursor and Copilot rule files for you and needs sign-in [11] |
| Prompt library | Yes, with {{variables}} and version history; installs prompts as slash commands in Claude Code, Codex, OpenCode, Antigravity and Qwen Code | Not listed [1] | Not listed [5] | No library; an open-source skill turns your history into agent skills [9] |
| Resume a session | One click reopens Claude Code, Codex and OpenCode sessions; editor chats open the project in Cursor or VS Code | Copies the resume command or resumes in a terminal [1] | Copies the command or opens it in Terminal, iTerm2 or Warp [5] | Yes, even into a different agent, and on another machine through its cloud [10] |
| Account and network | No account. Nothing is uploaded; the only network call is the Pro licence check | No account; the desktop app uploads nothing and has no telemetry [1] | No account, no telemetry; it checks for updates [5] | No account for local saving; cloud sync and sharing need sign-in. Usage analytics are on by default and can be turned off [12] [14] |
| Open source | No | Yes, MIT [1] | Yes, MIT [5] | Command-line tool yes (Apache-2.0); editor extensions no [9] |
| Price | Free tier with limits; Pro is $29.99 a year or $3.99 a month | Free [1] | Free [5] | Extensions and command-line tool free; the cloud has a free plan [15] |
Where the others do better
Claude Code History Viewer is free and open source, reads 31 assistants to Prompt Deck's 11, and runs on Windows and Linux as well as the Mac. It also shows cost analytics and edits Claude Code's own settings.
Agent Sessions is free, open source and just as native. It reads agents Prompt Deck doesn't, such as Cline and Devin, shows a live Codex and Claude usage meter, and can resume a session straight into iTerm2 or Warp.
SpecStory works on any operating system and saves your history as Markdown files you can commit to git. It can resume a session in a different agent or on another machine, and it has team sharing through its cloud.
Where Prompt Deck is different
None of the other three keeps a prompt library with variables that installs as slash commands in your agents. None of them puts CLAUDE.md, AGENTS.md, GEMINI.md and the other memory files on one screen where you can edit them, or append the same note to several at once. If you only want to read old sessions, the free tools above may be all you need. If you also reuse prompts and keep memory files in step across tools, that is what Prompt Deck is for.
Sources
- Claude Code History Viewer on GitHub
- Claude Code History Viewer changelog
- Claude Code History Viewer settings manager
- Claude Code History Viewer app config (minimum macOS)
- Agent Sessions on GitHub
- Agent Sessions changelog
- Agent Sessions website
- Agent Sessions Xcode project (minimum macOS)
- SpecStory on GitHub
- SpecStory docs: resume
- SpecStory docs: features
- SpecStory docs: privacy
- SpecStory docs: cloud search
- SpecStory docs: cloud overview
- SpecStory product summary (llms.txt)
Spotted something out of date? Tell me and I will fix it.