SpaceXAI Open-Sources Grok Build: Rust Agent Harness, TUI, and Tool Layer Behind Its CLI Code

admin 3 hours ago

0 0 5 minutes read

SpaceXAI Open-Sources Grok Build: Rust Agent Harness, TUI, and Tool Layer Behind Its CLI Code

SpaceXAI has open sourced the Grok Buildthe terminal-based AI code agent behind it grok The CLI. The source hit GitHub today. Release includes agent harness, TUI, CLI shell, and developer tools under the Apache 2.0 license

What is Grok Build?

A harnesses scaffolding around the model. It compiles the context, calls the model, parses the response, and dispatches the tool calls.

Grok Build launched as an early beta on May 25, 2026. An agent that understands your codebase, edits files, executes shell commands, and searches the web. It also manages long-running tasks. It works like a full-screen, mouse-friendly TUI.

There are three areas. There is an interactive TUI, a headless mode for writing and CI. Editors have embedded it with Agent Client Protocol (ACP).

xai-grok-shell’} }; var OUT={ tui:{name:’Render and update’,desc:’TUI paints result: scrollback, inline diff, program update.’,list:[‘Every change arrives as a reviewable diff’,’Checkpoints let you roll the workspace back’,’Rendering and input handling are in the pager crate’]}, headless:{ name:’Write to stdout’,desc:’The final response is streamed as a text or JSON stream.’,list:[‘Pipe it into another process or a CI check’,’No terminal UI is required’,’Exit code carries the outcome’]}, acp:{name:’Return ACP frames’,desc:’Results are returned to the host application as protocol messages.’,list:[‘The host decides how to display the diff’,’Useful for editor integrations and bots’,’Same tool results, structured transport’]} }; var STEPS=[ {id:’entry’,x:20,y:56,crate:’crates/codegen/xai-grok-pager’}, {id:’ctx’,x:250,y:56,crate:’crates/codegen/xai-grok-shell’,name:’Assemble context’,desc:’The harness discovers what this directory gives the model.’, list:[‘AGENTS.md, skills, plugins, hooks, MCP servers’,’config.toml picks the model and base_url’,’grok inspect prints exactly what was found’],log:’ctx found AGENTS.md, 3 capabilities, 1 MCP server’}, {id:’model’,x:480,y:56,crate:’model endpoint (api.x.ai or your own)’,name:’Call model’,desc:’Combined core goes to grok-4.5, or any model, edit list:[‘Default binary authenticates in the browser’,’XAI_API_KEY works in non-browser environments’,’A local base_url keeps the whole loop on your machine’],log:’model episode – 4.5 -> 1 tool call‘}, {id:’parse’,x:710,y:56,crate:’crates/codegen/xai-grok-shell’,name:’Parse the response’,desc:’The runtime separates the prose from the tool calls and validates the arguments.’, list:[‘Text goes to the transcript’,’Tool calls go to the dispatcher’,’Malformed calls are rejected, not executed’]}, {id:’tools’,x:710,y:216,crate:’crates/codegen/xai-grok-tools’,name:’Dispatch tools’,desc:’Tool usage: terminal, file editing, search, and web.’, list:[‘Read and edit files across the repo’,’Run shell commands’,’Search the codebase and the web’]}, {id:’work’,x:480,y:216,crate:’crates/codegen/xai-grok-workspace’,name:’Touch workspace’,desc:’The workspace layer manages the file system, VCS, operations, and test environments.’, list:[‘Every edit is checkpointed for rollback’,’Git state is tracked as the agent works’,’Sandboxing config lives in the CLI crate closure’]}, {id:’gate’,x:250,y:216,crate:’crates/codegen/xai-grok-shell’,name:’Loop or finish’,desc:’The results of the tool re-enter the context. Multiple tool calls imply another pass.’, list:[‘Results are appended, then the model runs again’,’The loop exits when no tool call comes back’,’Long tasks can run many passes’]}, {id:’out’,x:20,y:216,crate:’crates/codegen/xai-grok-pager’} ]; var W=170,H=64; function between(s){ return {x:s.x+W/2,y:s.y+H/2};} var PATHS=[ ‘M190 88 L246 88’, // entry -> ctx ‘M420 88 L476 88’, // ctx -> model ‘M650 88 L706 88’, // model -> parse ‘M795 120 L795 212’, // parse -> tools ‘M706 248 L654 248’, // tools -> work ‘M476 248 L424 248’, // work -> gate ‘M335 216 L335 124’, // gate -> ctx (loop up) ‘M246 248 L194 248’ // gate -> out ]; function el(t,a){var e=document.createElementNS(NS,t);for(var k in a)e.setAttribute(k,a[k]);return e;} var wireEls=[]; PATHS.forEach(function(d,i){ var p=el(‘path’,{d:d,’class’:’wire’,’marker-end’:’url(#gbA)’}); WIRES.appendChild(p); wireEls.push(p); }); // loop label var lt=el(‘text’,{x:342,y:172,’class’:’crate’}); lt.textContent=”loop while tool_calls”; WIRES.appendChild(lt); var boxEls={},labEls={}; STEPS.forEach(function(functions,i){ var g=el(‘g’,{}); var r=el(‘rect’,{x:sx,y:sy,width:W,height:H,rx:2,’class’:’nbox’}); var bar=el(‘rect’,{x,height:DSwidth:4,H:H[i%BANDS.length]}); var t1=el(‘text’,{x:s.x+14,y:s.y+26,’class’:’nlab’}); var t2=el(‘text’,{x:s.x+14,y:s.y+45,’class’:’crate’}); g.appendChild(r);g.appendChild(bar);g.appendChild(t1);g.appendChild(t2); NODES.appendChild(g); boxEls[s.id]=r; labels[s.id]={t:t1,c:t2}; }); function shortCrate(c){ var p=c.split(‘/’); return p[p.length-1];} function paintLabels(){ STEPS.forEach(function(s){ var nm=s.name, cr=s.crate; if(s.id===’entry’){nm=ENTRY[mode].name;} if(s.id===’out’){nm=OUT[mode].name;} labels[s.id].t.textContent=nm; labels[s.id].c.textContent=shortCrate(cr); }); } paintLabels(); var LOG=root.querySelector(‘#gb-log’); function login(html){ LOG.innerHTML+=’n’+html; LOG.scrollTop=LOG.scrollHeight; } function resetLog(){ LOG.innerHTML=’<code>'+ENTRY[mode].log+'n> Add retry logic to fetch_user() and run the tests.</code>‘; } var TOOLPASS=[ {t:’tool search “fn fetch_user”‘,w:’work read src/api/user.rs (1 match)‘}, {t:’tool edit src/api/user.rs +14 -2′,w:’work checkpoint written (rollback available)‘}, {t:’tool terminal cargo test -p api‘,w:’work 12 passed 0 failed‘} ]; var side={crate:root.querySelector(‘#gb-crate’), name:root.querySelector(‘#gb-name’), desc:root.querySelector(‘#gb-desc’), list:root.querySelector(‘#gb-list’)}; var AG=root.querySelector(‘#gb-agents’); var fills=AG.querySelectorAll(‘.fill’); function showSide(s){ var nm=s.name,ds=s.desc,ls=s.list; if(s.id===’entry’){nm=ENTRY[mode].name;ds=ENTRY[mode].desc;ls=ENTRY[mode].list;} if(s.id===’out’){nm=OUT[mode].name;ds=OUT[mode].desc;ls=OUT[mode].list;} side.crate.textContent=s.crate; side.name.textContent=nm; side.desc.textContent=ds; side.list.innerHTML=”; (ls||[]).forEach(function(x){var li=document.createElement(‘li’);li.textContent=x;side.list.appendChild(li);}); AG.style.display=(subs&(s.id===’tools’||s.id===’work’))?’block’:’none’; if(subs&&s.id===’tools’){ fill[0].style.width=”0%”;fills[1].style.width=”0%”;fills[2].style.width=”0%”; setTimeout(function(){fills[0].style.width=”90%”;fills[1].style.width=”65%”;fills[2].style.width=”40%”;},60); } } function litWires(i){ wireEls.forEach(function(p,j){ var on=false; if(i>0&&&i<7) on=(j===i-1); uma(i===1&&pass>1) on=(j===0)||(j===6); if(i=7=j=7); if(i=7); p.setAttribute(‘class’, open?wire lit’:’wire’); p.setAttribute(‘marker-end’,on?’url(#gbB)’:’url(#gbA)’); } function highlight(i){ STEPS.forEach(function(s,j){ boxEls[s.id].setAttribute(‘class’,j===i?’box open’:’box’);}); DOT.setAttribute(‘r’,’7′); DOT.setAttribute(‘fill’,BANDS[i%BANDS.length]); var c=mid(STEPS[i]); DOT.setAttribute(‘cx’,cx);DOT.setAttribute(‘cy’,cy); litWires(i); showSide(STEPS[i]); } function reduced(){ return window.matchMedia&&window.matchMedia(‘(prefers-reduced-motion: reduce)’).matches; } function move(from,to,cb){ var a=mid(STEPS)[from]),b=between(STEPS[to]); if(reduced()){cb();return;} var t0=null,dur=520; function frame(ts){ if(!t0)t0=ts; var k=Math.min(1,(ts-t0)/dur), e=k<.5?2*k*k:-1+(4-2*k)*k; DOT.setAttribute('cx',a.x+(bx-ax)*e); DOT.setAttribute('cy',a.y+(by-ay)*e); if(k<1){raf=requestAnimationFrame(frame);}else{cb();} } raf=requestAnimationFrame(frame); } function emit(i){ var s=STEPS[i]; if(i===0){return;} if(s.id==='ctx'){logline(pass===1?s.log:'ctx tool effects added (pass ‘+pass+’)‘);} else if(s.id===’model’){ logline(pass<3?s.log:'model episode – 4.5 -> no tool calls, made‘);} else if(s.id===’tools’){ logline(TOOLPASS[Math.min(pass-1,2)].t);if(subs)logline(‘a tool work -> 3 subagents, windows own context‘);} else if(s.id===’work’){ logline(TOOLPASS[Math.min(pass-1,2)].w);} else if(s.id===’gate’){logline(pass<3?'loop tool call returned, another pass’:’loop out’);} else if(s.id===’out’){ logline(‘done 1 file changed, passed 12 tests’);} } function next(){ if(idx<0){pass=1;idx=0;highlight(0);root.querySelector('#gb-it').textContent=pass;return;} var from=idx,to; if(idx===6){ to=(pass<3)?1:7; if(to===1){pass++;root.querySelector('#gb-it').textContent=Math.min(pass,3);} } else if(idx===7){ stop(); reset(); come back; } else { to=idx+1; } idx=to; travel(from,to,function(){ highlight(to);out(to);if(playing)setTimeout(function(){if(playing) next();},680);}); } function stop(){playing=false;root.querySelector('#gb-play').textContent="Play";if(raf)cancelAnimationFrame(raf);} function reset(){stop();idx=-1;pass=0;root.querySelector('#gb-it'";textContent=").textContent="). STEPS.forEach(functions){ boxEls[s.id].setAttribute('class','inbox'); }); wireEls.forEach(function(p){p.setAttribute('class','wire');p.setAttribute('marker-end','url(#gbA)');}); DOT.setAttribute('r','0'); AG.style.display='none'; side.crate.textContent="crates/codegen/xai-grok-pager"; side.name.textContent="Ready"; side.desc.textContent="Select a mode, then press Play. The loop repeats while the model keeps returning tool calls."; side.list.innerHTML=''; resetLog(); } root.querySelector('#gb-play').addEventListener('click',function(){ if(playing){stop();return;} playing=true;this.textContent="Pause"; if(idx<0){okulandelayo();emit(0);setTimeout(umsebenzi(){if(idlala)okulandelayo();},600);}okunye{okulandelayo();} }); root.querySelector('#gb-step').addEventListener('chofoza',umsebenzi(){stop();okulandelayo();}); root.querySelector('#gb-reset').addEventListener('chofoza',setha kabusha); root.querySelector('#gb-sub').addEventListener('chofoza',function(){ subs=!subs;this.setAttribute('aria-pressed',subs?'true':'false'); uma(idx>=0)showSide(STEPS[idx]); }); root.querySelectorAll(‘.gb-seg button’).forEach(function(b){ b.addEventListener(‘click’,function(){ root.querySelectorAll(‘.gb-seg button’).forEach(function(x){x.setAttribute(‘aria);-pressed’}}; b.setAttribute(‘aria-pressed’,’true’);mode=b.dataset.mode;paintLabels(); reset(}); resetLog(); function post(){ var h=root.offsetHeight+40; window.parent.postMessage({mtp-gbloop’}; } new ResizeObserver(root); setTimeout(post,120);

“>

What does the published site contain?

Building on that framework, SpaceXAI lists four published areas. I agent loop includes context binding, response parsing, and tool call dispatch. I tools covers how the agent reads, edits, and searches for code. I Terminal UI it includes rendering, input management, program updating, and a separate inline viewer. I expansion system includes skills, plugins, hooks, MCP servers, and subagents.

Those places point to named crates:

The way	Content
`crates/codegen/xai-grok-pager-bin`	The root-compile package; build i `xai-grok-pager` binary
`crates/codegen/xai-grok-pager`	TUI: scrollback, fast, modals, rendering
`crates/codegen/xai-grok-shell`	Agent runtime and leader/stdio/headless entry points
`crates/codegen/xai-grok-tools`	Use of tools (terminal, file editing, search)
`crates/codegen/xai-grok-workspace`	Host file system, VCS, execution, checkpoints
`third_party/`	Upstream fountain for sale (Mermaid painting stack)

Study that table as a study plan. Start at xai-grok-shell in the loop, then xai-grok-tools. One construction note is easy to remember. The root Cargo.toml is generated, and the README says to treat it as read-only.

How does the local-first approach work?

Apart from the test, SpaceXAI sets one valid result. Grok Build is now fully functional in the first place. Assemble it yourself, point it in the right direction, and remove everything config.toml.

# ~/.grok/config.toml  (Windows: %USERPROFILE%.grokconfig.toml)
[model.my-model]
model = "model-id"
base_url = "
name = "Display Name"
env_key = "API_KEY"

[models]
default = "my-model"

grok inspect then prints what the harness finds in the current directory. That includes maintenance sources, instructions, skills, plugins, hooks, and MCP servers.

How does the Grok Build compare?

Agent	First person license	Fork and fix	Model selection	External PRs
Grok Build (`xai-org/grok-build`)	Apache 2.0	It is allowed	Anywhere, with `config.toml`	Not accepted
Codex CLI (`openai/codex`)	Apache 2.0	It is allowed	OpenAI models	Open the PR line
OpenCode (`anomalyco/opencode`)	MIT	It is allowed	75+ suppliers	Community project
Claude Code	Ownership	Not given	Anthropic models	n/a

Use cases and examples

Given all that, four uses still exist today.

Check before you get it: Read xai-grok-tools before allowing the agent to run commands on the managed repo.
Inner harness fork: Apache 2.0 allows it; upstream integration is not provided.
Running with air gaps: Assemble in place, set base_url go to the inner end, then jump api.x.ai.
CI automation: Headless mode feeds programmed output into a pipeline step.

# Prebuilt binary (macOS, Linux, Windows)
curl -fsSL  | bash
grok --version

# Or build from source (needs protoc and the pinned Rust toolchain)
cargo build -p xai-grok-pager-bin --release   # -> target/release/xai-grok-pager

# Audit path: per-crate commands, because full-workspace builds are slow
cargo check -p xai-grok-tools
cargo test -p xai-grok-config
cargo clippy -p xai-grok-shell

# Headless run with machine-readable output
grok -p "Explain the architecture" --output-format streaming-json

# Route one run through a model declared in ~/.grok/config.toml
grok inspect
grok -p "Hello" -m my-model

Check out the GitHub Repo, announcement, and documentation.

Michal Sutter is a data science expert with a Master of Science in Data Science from the University of Padova. With a strong foundation in statistical analysis, machine learning, and data engineering, Michal excels at turning complex data sets into actionable insights.

admin 3 hours ago

0 0 5 minutes read