Bidwright
What it is
Bidwright is an open-source construction estimating workspace that manages bids from intake and takeoff through pricing, review, and quote delivery. Estimators can work manually, ask AI to review an estimate, or have an agent draft line items for review. PDF takeoff supports count, linear, area, and region annotations; 3D model takeoff extracts element quantities and supports filtered bills of materials and revision comparisons. The workspace lists viewers for formats including PDF, DOCX, XLSX, DWG/DXF, email files, Bluebeam markups, RTF, and ZIP. Its estimator library includes resources, ratebooks, assemblies, labor units, cost intelligence, playbooks, knowledge, and datasets. Desktop data is stored locally without a cloud account requirement, and manual workflows can run without outbound AI calls. Docker deployments can run offline or air-gapped after setup. The desktop app and Docker self-host plan are free; AI is disabled in the desktop app until users add their own key, and AI usage is billed by the provider. Hosted service is coming soon, with pricing not yet listed.
For and against
For
- Free desktop app for macOS, Windows, and Linux.
- PDF takeoff supports count, linear, area, and region marks.
- 3D model takeoff extracts element quantities.
- Manual workflows can run without AI calls.
- Desktop data is stored locally.
Against
- Desktop AI requires users to add their own key.
- AI usage is billed by the provider.
- Docker users manage their own infrastructure.
- Hosted service is coming soon.
Who it suits
The desktop app is positioned for a single estimator, owner-operator, or evaluator. Docker self-host is positioned for a small team or power user prepared to manage infrastructure.
Getting started
- Open the Bidwright website
- Choose the desktop app for macOS, Windows, or Linux, or choose Docker self-hosting
- For desktop AI, add a key from a supported provider or use a local LM Studio or Ollama option
- For Docker, set up and manage your own infrastructure
Limits to know
The free desktop app is single-user, and AI is disabled until you add your own key. Docker self-hosting is multi-user but requires self-managed infrastructure; AI use is billed directly by your provider.
The verdict
Bidwright offers manual and AI-assisted estimating workflows, with PDF and 3D model takeoff options. Its free desktop and self-host choices come with different setup and support responsibilities, and hosted availability is still pending.
Questions
Is Bidwright free?
The desktop app and Docker self-host plan are each listed at no cost.
What platforms does it support?
The listed platforms are API, Linux, macOS, self-hosted, web, and Windows. The desktop plan is for macOS, Windows, and Linux.
Does the desktop app include AI?
AI is disabled until you add your own key. Supported providers include Anthropic, OpenAI, OpenRouter, and Gemini, with local LM Studio and Ollama options also described.
How is AI usage billed?
AI usage is billed directly by your provider. Bidwright says it does not proxy or rebill tokens.
Can Bidwright run offline?
Manual workflows can run without outbound AI calls. Docker deployments can run offline or air-gapped after setup.
What does the AGPL license mean for embedding Bidwright in another SaaS?
The pricing FAQ says that embedding Bidwright in another SaaS under AGPL requires offering source to users. It directs users seeking another arrangement to ask about a commercial license.
The long read
Overview
Bidwright is an open-source construction estimating workspace for handling a bid from intake and takeoff through pricing, review, and quote delivery. Its scope includes digital takeoff, bid management, and labor and material costs, with an estimator library for organizing resources such as ratebooks, assemblies, labor units, and datasets.
It offers both manual estimating and AI-assisted workflows: an estimator can ask AI to review an estimate or have an agent draft line items for human review. AI is optional rather than a requirement for manual work.
For a broader look at this category, see Painting and construction estimating software.
Key features
Takeoff and documents
PDF takeoff supports count, linear, area, and region annotations. For 3D models, Bidwright extracts element quantities and supports filtered bills of materials and revision differences. The workspace lists viewers for PDF, DOCX, XLSX, DWG/DXF, MSG/EML, Bluebeam markups, RTF, and ZIP files.
Estimating resources and AI
The estimator library covers resources, ratebooks, assemblies, labor units, cost intelligence, estimator playbooks, knowledge, and datasets. AI connections use the user's own provider keys for Anthropic, OpenAI, OpenRouter, or Gemini; local options through LM Studio and Ollama are also described. AI usage is billed directly by the provider, not through Bidwright.
Manual workflows can run without outbound AI calls. The site says Docker deployments can operate offline or air-gapped after setup. The desktop app stores data locally and does not require a cloud account; Bidwright says it cannot see organization AI keys.
Pricing
Bidwright is free, with two listed options at no cost:
- Desktop app: Free. Single-user for macOS, Windows, and Linux; AI is disabled until you add your own key.
- Docker self-host: Free. AGPL-3.0, multi-user server, and self-managed infrastructure.
The Hosted option is coming soon, with a waitlist open. Its price is not listed; pricing is to be published before general availability.
The pricing FAQ says that embedding Bidwright in another SaaS under AGPL requires offering source to users. It directs anyone seeking a different arrangement to ask about a commercial license.
Platforms
Bidwright lists API, Linux, macOS, self-hosted, web, and Windows platforms, and describes its overall platform as hybrid. The desktop app is intended for a single user, while Docker self-hosting provides a multi-user server. Hosted service is not yet generally available.
Who it's for
The desktop app is positioned for one estimator, an owner-operator, or someone evaluating the software. Docker is aimed at a small team or a power user comfortable managing infrastructure. It may suit estimators who want to keep manual work independent of AI, or who prefer connecting their own AI provider or local model options.
Pros and cons
Pros
- Free desktop and Docker self-host options are listed.
- PDF and 3D model takeoff cover several annotation and quantity workflows.
- Manual estimating does not depend on outbound AI calls.
- Users can choose provider-key or local-model AI integrations.
- The desktop app stores data locally and does not require a cloud account.
Cons
- Desktop AI requires users to supply their own key, and provider usage is billed separately by that provider.
- Docker self-hosting requires self-managed infrastructure; support is through GitHub Issues.
- Hosted service, pricing, SSO, audit logs, and data residency are not available yet; the latter capabilities are planned.
- Using the software in another SaaS under AGPL comes with a source-offering requirement unless a commercial arrangement is made.
Alternatives
Other construction and estimating options include OpenConstructionERP, Trimble Viewpoint Vista, Trimble Quadri, CostX, Buildxact, HCSS, QuoteSoft, and eTakeoff.
Verdict
Bidwright is a free, open-source estimating workspace with a wide takeoff and document scope, estimator resources, and optional AI workflows. Its fit depends on the working setup: the desktop app keeps data local for an individual, while Docker supports multi-user use but puts infrastructure management on the user. Hosted availability and pricing remain unresolved, so teams seeking a ready hosted deployment should account for that limitation.
Facts on the record
- 2D takeoff
- PDF takeoff supports calibrated count, linear, area and region annotations, symbol detection, region questions and PDF markup export.bidwright.app · read 14 Oct 2026
- 3D takeoff
- 3D workflows ingest IFC, Revit through Autodesk APS, DWG and Navisworks files, extract element quantities, create filtered BOMs and compare revisions.bidwright.app · read 14 Oct 2026
- AI billing
- AI usage is billed directly by the user's provider; Bidwright says it does not proxy or rebill tokens.bidwright.app · read 30 Sept 2026
- AI intake
- The agent can classify bid-package documents, create worksheets, take off symbols and draft cited line items while the user remains editor of record.bidwright.app · read 14 Oct 2026
- AI integrations
- Bidwright accepts user keys for Anthropic, OpenAI, OpenRouter and Gemini, or local LM Studio with Ollama embeddings.bidwright.app · read 14 Oct 2026
- AI review
- Review combines rule-based quality checks with cited AI coverage, gap and risk findings that users can resolve, dismiss or defer.bidwright.app · read 14 Oct 2026
- AI workflows
- Users can estimate manually, have AI review an estimate, or let an agent draft line items for their review.bidwright.app · read 30 Sept 2026
- API and MCP
- The public repository describes a Fastify API, CLI endpoints under /api/cli/* and an MCP server exposing estimate, knowledge, model, quote, review, system and vision tools.github.com · read 14 Oct 2026
- Data and security
- The desktop app stores data locally and does not require a cloud account; the site says organization AI keys are not visible to Bidwright.bidwright.app · read 30 Sept 2026
- Data control
- The site says desktop data stays on the machine, no cloud account is required, and local deployments can operate offline or air-gapped.bidwright.app · read 14 Oct 2026
- Desktop packaging
- Desktop releases are Electron bundles with embedded Postgres, API and web app for macOS DMG, Windows installer and Linux AppImage.bidwright.app · read 14 Oct 2026
- Document support
- The workspace lists viewers for PDF, DOCX, XLSX, DWG/DXF, MSG/EML, Bluebeam markups, RTF, and ZIP files.bidwright.app · read 30 Sept 2026
- Document viewers
- The workspace lists viewers for PDF, DOCX, XLSX, DWG/DXF, MSG/EML, Bluebeam markups, RTF and ZIP files.bidwright.app · read 14 Oct 2026
- Documentation depth
- The repository includes setup documentation, Docker files, environment examples, API/web/worker apps and a monorepo development stack using Node.js 20+, pnpm 10+ and Docker Desktop.github.com · read 14 Oct 2026
- Estimating library
- The library contains resources, ratebooks, assemblies, tiered labor units, cost intelligence, estimator playbooks, knowledge and datasets.bidwright.app · read 14 Oct 2026
- Estimator library
- The library includes resources, ratebooks, assemblies, labor units, cost intelligence, estimator playbooks, knowledge, and datasets.bidwright.app · read 30 Sept 2026
- Extensibility
- The repository documents a plugin framework with auditable extensions plus provider adapters for several hosted and local models.github.com · read 14 Oct 2026
- Hosted availability
- Managed hosting is still a waitlist product and its pricing has not yet been published.bidwright.app · read 14 Oct 2026
- Hosted controls
- Hosted is coming soon; SSO, audit logs, and data residency are listed as planned capabilities.bidwright.app · read 30 Sept 2026
- Intended users
- The desktop app is positioned for one estimator, an owner-operator, or an evaluator, while Docker is positioned for a small team or power user.bidwright.app · read 30 Sept 2026
- Key handling
- Bidwright says organizations bring their own provider keys, the service can see usage but not keys, and providers bill AI usage directly.bidwright.app · read 14 Oct 2026
- License
- Source code is licensed AGPL-3.0-only; embedding it in another SaaS requires offering source to users, with commercial licensing available for different terms.github.com · read 14 Oct 2026
- License limit
- The pricing FAQ says embedding Bidwright in another SaaS under AGPL requires offering source to users, and directs users seeking another arrangement to ask about a commercial license.bidwright.app · read 30 Sept 2026
- Limit: demo reliability
- The getting-started guide warns that the hosted demo can return HTTP 502 errors and may be unreliable.github.com · read 14 Oct 2026
- Limit: installer signing
- Release 0.2.2 notes say macOS and Windows installers are unsigned even though builds and checks passed.github.com · read 14 Oct 2026
- Limit: Intel Mac startup bug
- Earlier release notes document an Intel Mac startup crash linked to a missing embedded-Postgres platform component.github.com · read 14 Oct 2026
- Limit: Windows relaunch bug
- Release notes for version 0.2.1 describe a Windows relaunch crash caused by timezone resolution and leftover PostgreSQL or Node locks.github.com · read 14 Oct 2026
- Maker
- The repository is owned by braedonsaunders and the license identifies Braedon Saunders as copyright holder.github.com · read 14 Oct 2026
- Manual estimating
- The manual workflow includes a Snap 10-line quote and can run locally without outbound AI calls.bidwright.app · read 14 Oct 2026
- Model cost
- The pricing page estimates per-bid costs of $0.40–$1.20 for cached Claude Sonnet 4.6, $1.50–$4.00 for Claude Opus 4.7 review, $0.60–$1.80 for mixed GPT-5 and $0.30–$0.90 for Gemini 2.5 Pro.bidwright.app · read 14 Oct 2026
- Offline use
- Manual workflows can run without outbound AI calls, and the site says Docker deployments can run offline or air-gapped after setup.bidwright.app · read 30 Sept 2026
- Operations dashboard
- A shop dashboard tracks pipeline value, margins, win rate, aging quotes, project status and client profitability.bidwright.app · read 14 Oct 2026
- Product
- Bidwright is an open-source construction estimating workspace covering bid intake, takeoff, pricing, review and quote delivery.bidwright.app · read 14 Oct 2026
- Production sizing
- The deployment guide suggests 4 vCPU, 16 GB RAM and 80 GB SSD for a small team, with 8 vCPU, 32 GB RAM and 160 GB SSD for heavier concurrent use.bidwright.app · read 14 Oct 2026
- Release
- Version 0.2.2 was released on 8 October 2026 with estimating and reference-search updates, quote PDF email attachments and desktop builds.github.com · read 14 Oct 2026
- Repository activity
- The public GitHub repository shows 64 stars, 18 forks and more than 1,200 commits.github.com · read 14 Oct 2026
- Self-host stack
- The Docker installer starts a Next.js web app, Fastify REST API, BullMQ worker, PostgreSQL with pgvector, Redis and optional Ollama.bidwright.app · read 14 Oct 2026
- Support
- Docker self-host users are directed to community support through GitHub Issues and Discussions.bidwright.app · read 14 Oct 2026
- System requirements
- Docker Desktop 4.30 or newer, about 10 GB free disk and 8 GB recommended memory are listed for the self-host installer.bidwright.app · read 14 Oct 2026
- Takeoff
- PDF takeoff supports count, linear, area, and region annotations, while 3D model takeoff extracts element quantities and supports filtered BOMs and revision diffs.bidwright.app · read 30 Sept 2026