Most of what people sell as AI agents for general contractors is software with a demo call attached. This is not that. It is a working agent you install into the Claude or ChatGPT account you already pay for, from one click or one pasted line, and run on the next invitation to bid that lands in your inbox. We build agents like this for companies that have no engineers on payroll; this one we publish free, because the bid desk is where a contractor's week goes to die and it is the easiest place to prove what an agent can honestly do.
For operators · free agent
Free AI agent for general contractors
Turns bid documents into bid/no-bid sheets, scope letters, change orders and RFIs, without ever guessing a number.
- One paste turns an ITB or RFP into a bid/no-bid sheet: deadlines, bonds, retainage, liquidated damages and the traps, with every gap marked instead of guessed.
- It also drafts scope letters with real exclusions, change order emails that reserve your rights, and RFI lists, all in a plain contractor voice.
- It never invents a number, a date or a clause: anything your documents do not state comes back as NOT STATED with a list of what to chase.
- Installs into ChatGPT or Claude from one click or one pasted line: no download, no email gate, and nothing to configure.
Built by Liam Anderson at agentclaw
Install on ChatGPT
opens ChatGPT and sends the instruction
copies the instruction, opens Claude
Or paste it yourself, Claude Code included
Create this skill for me, exactly as its files specify: https://www.agentclawhq.com/agents/bid-response-agent-for-general-contractors/skill.md- Version
- v1.0
- Updated
- Aug 13, 2026
- Works with
- ChatGPT · Claude · Claude Code
- Price
- Free, no email gate
What it takes off your desk
01
Reads an ITB, RFP or bid invitation and produces a one-page bid summary: project, money, clock, eligibility, flags and the three questions worth asking.
02
Drafts scope letters and proposal cover letters with inclusions, exclusions and qualifications, and suggests the exclusions contractors commonly forget.
03
Writes change order emails and impact notices that give notice firmly, reserve cost and schedule rights, and never apologize for the change.
04
Builds numbered RFI lists where each question cites the sheet or spec section your documents give, flags NO REF when they do not, and says why the answer moves price or schedule.
05
Marks everything the documents do not state as NOT STATED and hands you the chase list, so a gap costs a phone call instead of a bid.
What a run looks like
example run
you paste
INVITATION TO BID: Cedar Ridge Public Works Facility, Phase 2 The City of Cedar Ridge invites sealed bids for the Phase 2 expansion of its Public Works Facility at 400 Industrial Parkway, Cedar Ridge: a 14,200 SF pre-engineered metal building addition with associated sitework, paving, and utility extensions. Sealed bids will be received at the City Clerk's office, 100 Main Street, until 2:00 p.m. local time, Thursday, September 10, 2026 [...] Each bid shall be accompanied by bid security in the amount of five percent (5%) of the base bid. The successful bidder will furnish Performance and Payment Bonds each in the amount of one hundred percent (100%) of the contract price. Liquidated damages will be assessed at $750 per calendar day beyond the contract time of 240 calendar days from Notice to Proceed. Retainage of 5% will be withheld from progress payments. Bids shall remain open for 90 days after opening. Contractors must be licensed in the State and prequalified with the City for projects over $2,000,000. This project is subject to state prevailing wage rates.
the agent returns
BID SUMMARY: Cedar Ridge Public Works Facility, Phase 2
MONEY
Bid security: 5% of base bid
P&P bonds: 100% of contract price each (Performance and Payment Bonds)
Retainage: 5% of progress payments (step-down NOT STATED)
Liquidated dmgs: $750 per calendar day beyond the 240-day contract time
(cap NOT STATED)
Payment terms: NOT STATED
FLAGS
- Liquidated damages of $750 per calendar day beyond 240 days, with no cap
on total LDs stated
- Bid must remain open 90 days after opening, past the 60-day pricing-hold
threshold
- Prequalification required above $2,000,000, with no engineer's estimate
given to check the project against that line
THREE QUESTIONS WORTH ASKING
1. What is the estimated contract value, and are we already prequalified
with the City of Cedar Ridge? The ITB triggers prequalification above
$2,000,000 but gives no engineer's estimate to check the project against.
2. Can our pre-engineered metal building and steel suppliers hold firm
pricing for the full 90 days the bid must stay open after the
September 10 opening?
3. No differing-site-conditions clause is stated even though utility
extensions are part of the scope: is subsurface or existing-utility
information available before we price it, or are we pricing that risk
blind?
[excerpt of the full summary: project, clock, eligibility and the ten-item
chase list are in the complete output]How installing works
Every path does the same thing: it hands your assistant one instruction pointing at this agent's source file, and the assistant reads the files and builds the skill in your own account. The ChatGPT button arrives with the instruction already submitted. The Claude button copies it to your clipboard and opens a new chat: paste and send. In Claude Code, or anywhere else, paste the same line from the panel. There is nothing to download and no account with us. If browsing is switched off on your plan, open the source link in your own browser instead: it is plain text, and copying it into a chat does the same job.
You need
- A ChatGPT or Claude account that can browse the web, which is how the one-line install fetches the agent's files.
- Your bid documents as text or PDF. When a PDF is a rough scan, paste the text instead.
The details
Who this is for
A general contractor or sub with a bid desk that is really one estimator, an office manager and too many tabs. You bid public and private work, the invitations arrive as forty-page PDFs, and the difference between a good week and a bad one is how fast someone reads them properly. Nobody on the payroll writes software, and nobody should have to: if you can paste a line into Claude or ChatGPT, you can run this.
If your shop already has a preconstruction department with its own software stack, this agent will still read an ITB faster than a person, but it was not built for you. It was built for the office where the estimator is also the project manager and, some weeks, the safety officer.
How it fits your estimating process
The agent drafts; your estimator decides. That split is built into its instructions, not left to luck. Paste an invitation to bid and it returns the bid/no-bid sheet: owner, delivery method, contract time, bonds, retainage, liquidated damages, the clock, the eligibility fine print, the flags, and the three questions this particular document raises. Every field it cannot find says NOT STATED, and the chase list at the bottom tells you exactly what to ask the owner before bid day.
Won the read and want to bid? The same agent drafts the scope letter with your inclusions and exclusions, and suggests the exclusions contractors commonly leave unsaid, marked as suggestions for you to confirm rather than decisions it made for you. During the job it writes change order emails that give notice the way notice should be given: firm, specific, rights reserved, no apology. And when the drawings do not add up, it turns your questions into a numbered RFI list the architect can actually answer.
What it deliberately does not do
It does not price. Unit costs, production rates, markup and risk money are estimator judgment, and an agent that guesses at them is a liability wearing a hard hat. Ask it to price a job and it will tell you, politely, that pricing stays with you.
It does not give legal advice. When a question turns on contract interpretation, bonding capacity or licensing, it says in one sentence that the question belongs to your attorney, broker or surety, then carries on with the drafting you asked for.
And it does not invent. No made-up bond percentages, no assumed deadlines, no clause numbers it did not see. The NOT STATED discipline is the whole point: a marked gap costs you a phone call, a confident guess costs you a bid.
What is inside the agent
Four plain-text files, and you can read every one of them before you trust them: the install source link on this page serves them as they are. SKILL.md is the agent itself: the four jobs, the hard rules, the voice. references/itb-checklist.md is the eighteen-point read it performs on every bid document. references/templates.md holds the exact output shapes, so Tuesday's bid summary looks like Monday's. README.md repeats the install instruction offline. That is the entire artifact: no installer, no telemetry, no account with us. It is the same file format we use on client work, packaged as a Claude Skill, and it is versioned, so when we improve it the page shows a new number and reinstalling takes a minute.
Fair questions
Is the example on this page real?+
Yes, with one honest caveat. The invitation to bid is a sample document we wrote for testing, and the output is an unedited excerpt from the agent's actual run on it. We do not publish client bid documents, for the reason you would hope.
Does it work on scanned PDFs?+
Usually. Claude and ChatGPT both read most construction PDFs directly, including decent scans. When a scan is rough, paste the text instead; the agent works from whatever text actually reaches it, and its summary tells you when sections came through incomplete.
Can it price the work?+
No, by design. It totals, checks arithmetic and points out inconsistencies, but unit costs and markup are estimator judgment. An agent that guesses at pricing is exactly the kind of AI a contractor should refuse to run.
What happens to our bid documents?+
They go where you paste them: your own Claude or ChatGPT account, under that platform's terms. Nothing is sent to us, ever; there is no account with us to send it to. The agent also reminds you, once and briefly, if you paste something marked confidential into a shared workspace.
What if it misses something?+
Treat the summary as a first pass with receipts, not a replacement for reading the contract you are about to sign. The agent marks what it could not find rather than papering over it, which makes its misses visible. On a bid you intend to win, the estimator still reads the document; they just start from a map.
Why is a paid agency giving this away?+
Because the library is our proof of work. Some firms install an agent and run it themselves forever, which is a fine outcome. Some decide they want agents built around their own workflows and run for them month after month, and that service is the business we are in.
More agents like this
Want it built around your workflow, and run for you?
This agent is the generic version. We also build agents around the way your company actually works, install them, and run them for you month after month.