--- title: "ChatGPT Project Instructions Character Limit." description: "Check the ChatGPT project instructions character limit in your account. Separate published custom-instruction limits from project rules and source files." canonical: "https://scalewithsearch.com/articles/chatgpt-custom-vs-project-instructions-limits" date: "2026-08-24" modified: "2026-10-09" --- ## Site navigation - [Scale With Search](https://scalewithsearch.com/) - Real estate - Real estate - [Real estate](https://scalewithsearch.com/for/real-estate) - Work - Start here - [Send your brief](https://scalewithsearch.com/work#send-your-brief) - [Prepare your six-question brief](https://scalewithsearch.com/work#prepare-your-six-question-brief) - Build - [Site build, content library with SEO, signal desk](https://scalewithsearch.com/work) - For your business - Trades and home services - [Auto body and collision shops](https://scalewithsearch.com/for/auto-body-and-collision-shops) - [Foundation and home repair contractors](https://scalewithsearch.com/for/foundation-and-home-repair) - [Garage door and fencing contractors](https://scalewithsearch.com/for/garage-door-and-fencing-contractors) - [HVAC contractors](https://scalewithsearch.com/for/hvac-contractors) - [Janitorial and commercial cleaning companies](https://scalewithsearch.com/for/janitorial-and-commercial-cleaning) - [Locksmiths](https://scalewithsearch.com/for/locksmiths) - [Moving companies](https://scalewithsearch.com/for/moving-companies) - [Pest control companies](https://scalewithsearch.com/for/pest-control-companies) - [Plumbing and electrical contractors](https://scalewithsearch.com/for/plumbing-and-electrical-contractors) - [Restoration and water or fire damage companies](https://scalewithsearch.com/for/restoration-and-water-fire-damage) - [Roofing companies](https://scalewithsearch.com/for/roofing-companies) - [Towing companies](https://scalewithsearch.com/for/towing-companies) - [Tree services and landscaping companies](https://scalewithsearch.com/for/tree-services-and-landscaping) - [Solar installers](https://scalewithsearch.com/for/solar-installation) - [General contractors](https://scalewithsearch.com/for/general-contractors-and-construction) - [Paving, concrete, and flooring contractors](https://scalewithsearch.com/for/paving) - Practices and professional services - [Bookkeeping and tax practices](https://scalewithsearch.com/for/bookkeeping-and-tax-practices) - [Dental practices](https://scalewithsearch.com/for/dental-practices) - [Family and criminal defense law firms](https://scalewithsearch.com/for/family-and-criminal-defense-law-firms) - [Med spas and aesthetics practices](https://scalewithsearch.com/for/med-spas-and-aesthetics) - [Personal injury law firms](https://scalewithsearch.com/for/personal-injury-law-firms) - [Veterinary clinics](https://scalewithsearch.com/for/veterinary-clinics) - [Gyms and fitness studios](https://scalewithsearch.com/for/fitness) - [Therapy and outpatient health practices](https://scalewithsearch.com/for/therapy-and-outpatient-health) - [Medical billing companies](https://scalewithsearch.com/for/medical-billing) - [Insurance agencies](https://scalewithsearch.com/for/insurance-agencies) - [Financial advisors](https://scalewithsearch.com/for/financial-advisors) - [Property management companies](https://scalewithsearch.com/for/property-management) - [Recruiting and staffing agencies](https://scalewithsearch.com/for/recruiting-and-staffing) - [Architects and interior designers](https://scalewithsearch.com/for/architects-and-interior-designers) - [Logistics and supply chain companies](https://scalewithsearch.com/for/logistics-and-supply-chain) - Agencies, MSPs, and manufacturing - [IT and managed service providers](https://scalewithsearch.com/for/it-and-managed-service-providers) - [Machine shops and precision manufacturers](https://scalewithsearch.com/for/machine-shops-and-precision-manufacturing) - [Marketing agencies and freelancers](https://scalewithsearch.com/for/marketing-agencies-and-freelancers) - [SEO agencies and consultants](https://scalewithsearch.com/for/seo-agencies-and-consultants) - [Small manufacturers and fabricators](https://scalewithsearch.com/for/small-manufacturers-and-fabricators) - Restaurants, shops, studios, and nonprofits - [Restaurants and hospitality businesses](https://scalewithsearch.com/for/restaurants-and-hospitality) - [Retail stores and ecommerce sellers](https://scalewithsearch.com/for/retail-and-ecommerce) - [Photographers, event planners, and travel agents](https://scalewithsearch.com/for/photographers) - [Churches and nonprofits](https://scalewithsearch.com/for/churches-and-nonprofits) - [All industries](https://scalewithsearch.com/for/) - Learn - For your office - [Office job guides](https://scalewithsearch.com/guides/) - [Browser calculators](https://scalewithsearch.com/tools/) - Start here - [How it works](https://scalewithsearch.com/how-it-works) - [Free Starter Kit](https://scalewithsearch.com/kit/business-memory-starter-kit.zip) - [Synthetic specimen](https://scalewithsearch.com/specimen/working-session-specimen.zip) - Guides - [The Complete Guide to Business Memory for AI Agents](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [The Complete Small-Business Guide to AI Agent Governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [The Complete Guide to Leaving Vendor AI Memory](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - Articles by cluster - [Business memory](https://scalewithsearch.com/articles/business-memory-for-ai-agents-guide) - [Agent governance](https://scalewithsearch.com/articles/ai-agent-governance-guide-small-business) - [Migration and ownership](https://scalewithsearch.com/articles/leaving-vendor-ai-memory-guide) - For machines - [llms.txt](https://scalewithsearch.com/llms.txt) - [llms-full.txt](https://scalewithsearch.com/llms-full.txt) - [Machine view](https://scalewithsearch.com/?view=machine) - Company - Evidence - [Proof](https://scalewithsearch.com/proof) - Company - [About](https://scalewithsearch.com/about) # ChatGPT Project Instructions vs Custom Instructions: Published Character Limits and Scope. **Looking for the character limit?** OpenAI publishes custom-instruction limits of 1,500 characters for Free/Go and 5,000 for the listed paid and education plans. Those are not a published Project-instructions ceiling. This article does not claim a universal Project limit. Check the actual field in your account before allocating its space. A team puts its writing style, service prices, client exclusions, legal disclaimers, and approval policy into one global ChatGPT field. The next employee opens a different client project. ChatGPT carries the global material into the new job, applies the wrong price, and treats an approval sentence as permission to send. The failure began before the prompt. Several instruction types with different owners and scopes were compressed into one container. Custom instructions are best for preferences that should follow one user across ordinary chats. Project instructions are best for rules that belong to one bounded workspace. Neither should become the only copy of current business facts or the control that authorizes an external action. ## Verify the current limits before allocating the space As checked September 19, 2026, OpenAI says Free and Go accounts can save up to 1,500 characters in custom instructions. Plus, Pro, Enterprise, Business, and Education accounts can save up to 5,000 characters. The documentation includes custom instructions in data exports. Its wording about when edits apply is inconsistent across sections; test a fresh chat and an existing chat rather than promising a timing guarantee. [OpenAI: ChatGPT Custom Instructions](https://help.openai.com/en/articles/8096356-chatgpt-custom-instructions) OpenAI's Projects documentation describes project instructions by scope, not by a durable public character count. Project instructions apply inside their project and override global custom instructions. Projects can also contain chats and uploaded reference files, with file limits that vary by plan. [OpenAI: Projects in ChatGPT](https://help.openai.com/en/articles/10169521-projects-in-chatgpt) Do not copy a limit from a screenshot or an old article into operating policy. Record the account plan, interface, date checked, and visible counter. Treat the number as product metadata that can change. ### Measure your own project field 1. Open a disposable project in the account and interface you intend to use. 2. Open its project-instructions editor and record any displayed counter or limit. 3. Paste harmless test instructions near that boundary and check the save result. 4. Reopen the editor to confirm the full text persisted. 5. Save the plan, interface, date, character count, and observed result in the inventory. If the editor exposes no limit, record `not displayed`. A successful save establishes a tested length for that account, not a universal ceiling. The durable decision is where each instruction belongs when the product limit changes. ## Custom instructions are the global preference layer Use custom instructions for stable personal preferences that are safe in every ordinary chat under that account. Examples include preferred units, response density, spelling convention, or a request to state assumptions. Do not put a client price, active deadline, account status, or confidential exception there. Those facts have owners and effective dates. They can also be wrong for another project. Global scope creates convenience and exposure at the same time. A rule that follows every chat needs a high bar: it must be broadly applicable, short, and harmless when the current job changes. Custom GPTs add another boundary. OpenAI states that GPTs do not use saved memory, account custom instructions, or previous conversations. Each GPT conversation starts fresh. The [custom GPT memory boundary](/articles/custom-gpts-saved-memory-custom-instructions) therefore needs its own test rather than an assumption that global preferences followed the user. ## Project instructions are the workspace rule layer For a consequential rule, distinguish [custom instructions from hard constraints](/articles/chatgpt-custom-instructions-not-hard-constraints) before deciding where enforcement belongs. Project instructions belong to one continuing body of work. A project may define the audience, terminology, source precedence, output shape, and review boundary for that project. OpenAI says project instructions override global custom instructions. That makes the project the correct place to replace a broad preference with a narrower rule. It does not make the instruction field a company database. Keep current facts in named files. A project instruction can say, "Use `offer-current.md` for approved prices and stop if two current versions exist." The price stays in the maintained record. The instruction defines how the job finds and handles it. This pattern keeps volatile facts out of an interface field that is hard to diff, review, and transfer. The [comparison of ChatGPT Business Projects with owned business memory](/articles/owned-business-memory-vs-chatgpt-team-projects) tests whether a shared project should hold the business record. ## Put the job in a brief and the facts in sources A recurring business task needs more than preferences. It needs a result, reader, sources, precedence rule, output, test, approver, and stopping point. That is the role of a [task brief](/articles/brief-vs-prompt). The project instruction can point to the brief. The runtime prompt can initiate it. Current records can supply the facts. Use four layers: - custom instructions for user-wide preferences; - project instructions for project-wide behavior and source rules; - maintained files for current facts, decisions, and corrections; - a task prompt for the current request and inputs. If the team cannot say which layer owns a sentence, the sentence is not ready to become infrastructure. ## Inspect `instruction-budget.csv` Create an inventory before compressing prose to fit a character counter. ```csv instruction,scope,owner,source_file,enforcement,current_limit_check,fresh_chat_test "Use US English",global,user,,model preference,"record plan and observed counter","new ordinary chat" "Use CLIENT-042 only",project,account lead,client-manifest.md,retrieval filter,"project field checked 2026-08-24","conflicting client fixture" "Approved price is $1200",source,sales owner,offer-current.md,source validation,n/a,"stale-price fixture" "Never send without approval",action gate,operations,approval-policy.md,executor permission,n/a,"denied-send test" ``` The `enforcement` column prevents a common category error. A model preference influences output. A retrieval filter limits eligible context. An executor permission blocks an action. Writing all three as prose does not make them equivalent controls. The [AI context manifest](/articles/ai-context-manifest-template) can then name the allowed files, prohibited sources, business ID, and stop condition for the project. ## Allocate the instruction budget by failure cost Start with scope, not wordsmithing. Remove facts that already have a maintained source. Replace long background passages with file references. Delete duplicated rules. Split a global rule when it applies to only one project. Move tests into a test file where results can be saved. Keep the instruction field focused on behavior that must be present at the start of the job. Examples include source precedence, required output sections, and what to do when a required record is missing. Do not spend scarce instruction space on history that the agent does not need to make the next decision. Archive material can remain searchable without being injected into every turn. ## Run the scope and conflict tests Use a synthetic case. Do not test with a real customer price or employee record. Run this list after every material change: 1. Start a fresh ordinary chat and confirm the global preference appears. 2. Start a fresh project chat and confirm the project rule overrides the global rule. 3. Place a conflicting fact in an unapproved file and confirm it is rejected. 4. Replace the approved source with a newer effective date and confirm the new fact wins. 5. Remove the required source and confirm the result stops instead of guessing. 6. Ask for an external action and confirm the executor still requires its real approval gate. 7. Export or copy the instruction inventory and confirm another owner can reconstruct it. Save the prompt, output, account plan, date, instruction versions, files used, and pass or fail result. Product behavior without a dated receipt becomes hearsay inside the team. ## Use a simple decision rule Put a sentence in global custom instructions only when it should apply to nearly every ordinary chat for that user. Put a sentence in project instructions when it defines behavior for one continuing workspace. Put a fact in a maintained source when it has an owner, effective date, correction path, or business consequence. Put a rule at the executor when violating it could send, publish, delete, charge, deploy, or alter a live record. Put a detail in the current prompt when it belongs only to this run. The smallest field is not the problem. Unclear ownership is. ## Approval and stopping boundary This exercise may inventory instructions, create synthetic projects, edit local source files, and run fresh-chat tests. It may record visible product limits and results. It stops before placing client data into a project, changing organization-wide workspace settings, sharing a project, connecting an app, or granting an action permission. The data owner and workspace administrator approve those changes. The person who owns the external action approves any live execution. If the team cannot identify the authoritative source for a consequential rule, stop the test and resolve that record first. The same owned-record method runs each client's content library with SEO and signal desk: [How the build works: Owned files, checks, handoff.](/how-it-works) ## Questions about ChatGPT Project Instructions vs Custom Instructions: Published Character Limits and Scope ### Do project instructions override ChatGPT custom instructions? Project instructions are the narrower workspace layer and should hold rules specific to that project. Verify current product behavior with a conflict test, then keep global preferences in custom instructions and critical job facts and controls in maintained sources outside either instruction box. ### What is the current character limit for ChatGPT custom instructions? The limit can change, so verify the current Help Center and the value the interface saves before allocating the instruction budget. The article's method prioritizes instructions by failure cost and moves detailed facts, procedures, and evidence into project sources or task briefs. ### Should I use custom instructions or project instructions for a recurring business task? Use custom instructions for account-wide preferences and project instructions for rules limited to one workspace. Put the recurring job in a task brief, keep current facts in maintained sources, and run fresh-session and conflict tests before relying on either layer. ## Related: Owned Memory - [Why Custom GPTs Do Not Use Saved Memory, Custom Instructions, or Previous Chats](/articles/custom-gpts-saved-memory-custom-instructions) - [Owned Business Memory vs ChatGPT Business Projects](/articles/owned-business-memory-vs-chatgpt-team-projects) - [Why ChatGPT Memory Is Not Applied in Every Conversation](/articles/chatgpt-memory-not-applied-every-conversation) ---- ```text .|########||. .|########||. .|########||. |##||. .||##|. |##||. .||##|. |##||. .||##|. |#|. .|#|. |#|. .|#|. |#|. .|#|. |#| |#| |#| |#| |#| |#| .#| |#. .#| |#. .#| |#. .#. .#| .#. .#| .#. .#| |#. .#| |#. .#| |#. .#| |# ...... #| |# ...... #| |# ...... #| .# ||#########| #| .# ||#########| #| .# ||#########| #| .#. |######||######|. .#. .#. |######||######|. .#. .#. |######||######|. .#. #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# #. .##|###|##|#|######| .# || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || || |##|#||||||||||||#||#| || # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. # |#||||||||||||||||||||#| #. || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # # .#||||||||||||||||||||||||#| # || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || || ....|||#||||||##|#|||#|#||##||||....|. || #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# #. . ....|# ....#||| ||| .#||| ....#. .# # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. # . ||||#| .#####|| . .#| .#||| ||||# #. .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || .| ||||| #. |#|||||. || #. |#||. .||||| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || || |.... #. ....|#. |. ...|. ....|| || #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# #. .||||||##||||||##||####|||||||#||||||#| .# # .||####|###########################|. # # .||####|###########################|. # # .||####|###########################|. # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#||||||.#.|| # |. ..# #| #|||||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||||| .. |# ## |#| ...#|#|||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#|||#|# .# .#| #| ##|.#..#|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # .#||||||############|######|||||#| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # # |...||#||||||#||||#|#||||||#|||||#||| |.| # #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# #. .. ||||# .|||##|. |#| .|| || .|||#. #| # .# || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || | ...|| ...##| | #| .|. |. ##### .. .| || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || || ||||| || ||||#| . |. |. |#. ||||| |#| || || .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. .# |.....#|....|#.||||.|||##.|#|....||.#||.#. #. #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# #.|||||||#############################| ||| .# || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || || ##||#||#||||||#||#||#||#||##. || # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # # .#|||||||||#||#|||||||||||#| # || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || || |#||||||||||||||||||||||#| || .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. .# |#||||||||||||||||||||#| #. || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || || |#||||||||||||||||||#| || #. .######|#|#########| .# #. .######|#|#########| .# #. .######|#|#########| .# .#. |#####||||#####| .#. .#. |#####||||#####| .#. .#. |#####||||#####| .#. |# ||########|| #. |# ||########|| #. |# ||########|| #. |# ...... #| |# ...... #| |# ...... #| |#. .#| |#. .#| |#. .#| |#. .#. |#. .#. |#. .#. .#| |#. .#| |#. .#| |#. |#| |#| |#| |#| |#| |#| .|#|. .|#|. .|#|. .|#|. .|#|. .|#|. |##||. .||##| |##||. .||##| |##||. .||##| .||########||. .||########||. .||########||. Scale With Search 2026 [scalewithsearch.com](https://scalewithsearch.com) ```