<div align="center">

jcode

The most RAM efficient harness <br> The most intelligent harness

<a href="https://trendshift.io/repositories/25042?utm_source=repository-badge&amp;utm_medium=badge&amp;utm_campaign=badge-repository-25042" target="_blank" rel="noopener noreferrer"><img src="https://trendshift.io/api/badge/repositories/25042" alt="1jehuang/jcode | Trendshift" width="250" height="55"></a>

<a href="https://github.com/1jehuang/jcode/releases/download/readme-assets/jcode-yc-launch.mp4"> <img src="https://github.com/1jehuang/jcode/releases/download/readme-assets/jcode-yc-launch.webp" alt="jcode YC launch video" width="800"> </a> <br>

Website · Docs · SDK · Benchmarks · Features · Install · Quick Start · Further Reading · Contributing

</div>

---

<div align="center">

Installation

</div>

# macOS & Linux
curl -fsSL https://jcode.sh/install | bash

# Windows 11 (PowerShell 5.1+)
irm https://jcode.sh/install.ps1 | iex

Need Homebrew, source builds, provider setup, or want an agent to set it up for you? Jump to detailed installation.

Updating

Run /update in the TUI to download the latest stable release in the background and reload with your session preserved. From a terminal, use jcode update, then restart the client. Both commands use the same update policy, including for dev builds.

Older or equal release versions are skipped. For a development build, Jcode also compares the running binary's Git commit with the release tag. Builds ahead of, identical to, or diverged from the release are preserved. If ancestry cannot be verified locally or through GitHub, the update stops rather than risking a downgrade. The displayed dev patch includes a commit-count offset, so it is not used as a release version comparison.

This is the default features.update_channel = "stable" behavior. An explicit "main" channel still opts into source-branch updates. Use /rebuild or the self-dev build workflow to rebuild your own checkout.

---

<div align="center">

Performance & Resource Efficiency

</div>

jcode is built to be as performant and resource efficient as possible. Every metric is optimized to the bone, which is important for scaling multi-session workflows. Here we sample a few metrics to show the difference: RAM usage and boot up.

RAM comparison

<div align="center"> <table> <tr> <td valign="top" align="center" width="50%"> <strong>1 active session</strong> <table> <thead> <tr> <th>Tool</th> <th>PSS</th> <th>Comparison</th> </tr> </thead> <tbody> <tr> <td><strong>jcode (local embedding off)</strong></td> <td align="right"><strong>27.8 MB</strong></td> <td align="right">baseline</td> </tr> <tr> <td><strong>jcode</strong></td> <td align="right"><strong>167.1 MB</strong></td> <td align="right"><strong>6.0× more RAM</strong></td> </tr> <tr> <td><strong>pi</strong></td> <td align="right"><strong>144.4 MB</strong></td> <td align="right"><strong>5.2× more RAM</strong></td> </tr> <tr> <td><strong>Codex CLI</strong></td> <td align="right"><strong>140.0 MB</strong></td> <td align="right"><strong>5.0× more RAM</strong></td> </tr> <tr> <td><strong>OpenCode</strong></td> <td align="right"><strong>371.5 MB</strong></td> <td align="right"><strong>13.4× more RAM</strong></td> </tr> <tr> <td><strong>GitHub Copilot CLI</strong></td> <td align="right"><strong>333.3 MB</strong></td> <td align="right"><strong>12.0× more RAM</strong></td> </tr> <tr> <td><strong>Cursor Agent</strong></td> <td align="right"><strong>214.9 MB</strong></td> <td align="right"><strong>7.7× more RAM</strong></td> </tr> <tr> <td><strong>Claude Code</strong></td> <td align="right"><strong>386.6 MB</strong></td> <td align="right"><strong>13.9× more RAM</strong></td> </tr> <tr> <td><strong>Antigravity CLI</strong></td> <td align="right"><strong>243.7 MB</strong></td> <td align="right"><strong>8.8× more RAM</strong></td> </tr> </tbody> </table> </td> <td width="24"></td> <td valign="top" align="center" width="50%"> <strong>10 active sessions</strong> <table> <thead> <tr> <th>Tool</th> <th>PSS</th> <th>Comparison</th> </tr> </thead> <tbody> <tr> <td><strong>jcode (local embedding off)</strong></td> <td align="right"><strong>117.0 MB</strong></td> <td align="right">baseline</td> </tr> <tr> <td><strong>jcode</strong></td> <td align="right"><strong>260.8 MB</strong></td> <td align="right"><strong>2.2× more RAM</strong></td> </tr> <tr> <td><strong>pi</strong></td> <td align="right"><strong>833.0 MB</strong></td> <td align="right"><strong>7.1× more RAM</strong></td> </tr> <tr> <td><strong>Codex CLI</strong></td> <td align="right"><strong>334.8 MB</strong></td> <td align="right"><strong>2.9× more RAM</strong></td> </tr> <tr> <td><strong>OpenCode</strong></td> <td align="right"><strong>3237.2 MB</strong></td> <td align="right"><strong>27.7× more RAM</strong></td> </tr> <tr> <td><strong>GitHub Copilot CLI</strong></td> <td align="right"><strong>1756.5 MB</strong></td> <td align="right"><strong>15.0× more RAM</strong></td> </tr> <tr> <td><strong>Cursor Agent</strong></td> <td align="right"><strong>1632.4 MB</strong></td> <td align="right"><strong>14.0× more RAM</strong></td> </tr> <tr> <td><strong>Claude Code</strong></td> <td align="right"><strong>2300.6 MB</strong></td> <td align="right"><strong>19.7× more RAM</strong></td> </tr> <tr> <td><strong>Antigravity CLI</strong></td> <td align="right"><strong>1021.2 MB</strong></td> <td align="right"><strong>8.7× more RAM</strong></td> </tr> </tbody> </table> </td> </tr> </table> </div>

Time to first frame

<div align="center">

| Tool | Time to first frame | Range | Comparison | |---|---:|---:|---:| | jcode | 14.0 ms | 10.1–19.3 ms | baseline | | Antigravity CLI | 383.5 ms | 363.1–415.4 ms | 27.4× slower | | pi | 590.7 ms | 369.6–934.8 ms | 42.2× slower | | Codex CLI | 882.8 ms | 742.3–1640.9 ms | 63.1× slower | | OpenCode | 1035.9 ms | 922.5–1104.4 ms | 74.0× slower | | GitHub Copilot CLI | 1518.6 ms | 1357.4–1826.8 ms | 108.5× slower | | Cursor Agent | 1949.7 ms | 1711.0–2104.8 ms | 139.3× slower | | Claude Code | 3436.9 ms | 2032.7–8927.2 ms | 245.5× slower |

</div>

Measured on this Linux machine across 10 interactive PTY launches.

Time to first input

(time until typed probe text appears on the rendered screen; Antigravity uses its internal input-ready log marker because the sign-in screen suppresses probe echo.)

<div align="center">

| Tool | Time to first input | Range | Comparison | |---|---:|---:|---:| | jcode | 48.7 ms | 30.3–62.7 ms | baseline | | Antigravity CLI | 383.7 ms | 363.4–415.7 ms | 7.9× slower | | pi | 596.4 ms | 373.9–955.2 ms | 12.2× slower | | Codex CLI | 905.8 ms | 760.1–1675.7 ms | 18.6× slower | | OpenCode | 1047.9 ms | 931.1–1116.9 ms | 21.5× slower | | GitHub Copilot CLI | 1583.4 ms | 1422.8–1880.0 ms | 32.5× slower | | Cursor Agent | 1978.7 ms | 1727.3–2130.0 ms | 40.6× slower | | Claude Code | 3512.8 ms | 2137.4–9002.0 ms | 72.2× slower |

</div>

Measured on this Linux machine across 10 interactive PTY launches. Antigravity CLI was unauthenticated for this run; its sign-in screen rendered normally and emitted an internal CLI ready for user input marker, but did not echo the typed probe.

Additional clients / memory scaling

<div align="center">

| Tool | Extra PSS per added session | Comparison | |---|---:|---:| | jcode (local embedding off) | ~9.9 MB | baseline | | jcode | ~10.4 MB | 1.1× more RAM | | pi | ~76.5 MB | 7.7× more RAM | | Codex CLI | ~21.6 MB | 2.2× more RAM | | OpenCode | ~318.4 MB | 32.2× more RAM | | GitHub Copilot CLI | ~158.1 MB | 16.0× more RAM | | Cursor Agent | ~157.5 MB | 15.9× more RAM | | Claude Code | ~212.7 MB | 21.5× more RAM | | Antigravity CLI | ~86.4 MB | 8.7× more RAM |

</div> versions tested for this corrected memory rerun:

• jcode v0.9.1888-dev (be386f2)
• pi 0.62.0
• codex-cli 0.120.0
• opencode 1.0.203
• GitHub Copilot CLI 1.0.24 for the 1-session rerun, GitHub Copilot CLI 1.0.27 for the 10-session rerun
• Cursor Agent 2026.04.08-a41fba1
• Claude Code 2.1.86 (Claude Code)
• Antigravity CLI 1.0.0

<div align="center"> <a href="https://github.com/1jehuang/jcode/releases/download/readme-assets/jcode-performance-demo.mp4"> <img src="https://github.com/1jehuang/jcode/releases/download/readme-assets/jcode-performance-demo.webp" alt="jcode performance demonstration" width="900"> </a> <p><em>jcode performance demonstration</em></p> </div>

---

Memory (Agent memory)

Jcode embeds each turn/response as a semantic vector. Every turn does queries a graph of memories to efficiently find related memory entries via a cosine similarity check. The embedding hits are fed into the conversation, or optionally uses a memory sideagent which verifies the memories are relevant, and potentially does more work for information retrieval before injecting into the conversation. This results in a human like memory system which allows the agent to automatically recall relevant information to the conversation without actively calling memory tools or being a token burner. ot To have memories which are retrieved, they must also be extracted and stored. Every so often (semantic drift, K turns since last extraction, session end, etc), memories are extracted via a memory sideagent, and put into the memory graph.

The harness also provides explicit memory tools to allow the agent to actively search or store the memory without relying on a passive background