A creator's Instagram account can contain years of work, but turning that content into a portfolio website isn't as simple as embedding a few posts.
I ran into this while building a website for a content creator whose work included filmmaking, event coverage, styling, and wedding content.
The initial website had a design, photographs, and information from the creator's CV. What it lacked was the context behind the work. I couldn't confidently identify some of the projects, explain how they came together, or connect certain pieces of content to the services the creator offered.
I used Spidra to collect structured information from the creator's public Instagram content and made that information available to my AI coding agent.
The result was a portfolio built around actual projects, captions, videos, services, and audience feedback rather than generic descriptions.
This article explains how that workflow came together and how you can apply the same approach when building portfolios, agency websites, creator platforms, and other content-driven applications.
Why building a website from Instagram content is harder than it looks
Instagram is designed around individual posts. A portfolio website needs to organize those posts into something more meaningful.
A filmmaker might publish a finished video, several behind-the-scenes clips, and a caption explaining the creative process. On Instagram, those are separate pieces of content. On a portfolio, they might belong to a single project.
The same applies to event coverage. A creator may publish content before an event, share footage while it is happening, and create a recap afterward.
If you simply embed all three posts in a gallery, visitors see the content but may miss the story.
For this project, I started with the creator's CV, notes, and an existing visual portfolio. These provided useful background information and assets, but they didn't contain everything published on Instagram.
Some media files had generic names. Others were available only as thumbnails or still images. The context connecting them to specific projects wasn't always clear.
That left the coding agent making assumptions about which files belonged together and how the projects should be described.
A better approach was to collect the available source information first.
How Spidra fits into the workflow
Spidra provides structured data from public Instagram profiles, posts, and reels. It also offers ready-made Templates for common data collection tasks, including retrieving recent posts and post comments.
This means you don't have to build a custom scraping script for every step. For the portfolio, I used two complementary capabilities:
- Spidra Templates helped collect groups of posts and comments. A template accepts the inputs required for a particular task and returns structured records that can be reviewed or used in code.
- Spidra Scrape helped retrieve more detailed information from selected individual posts and reels, including available media and engagement metadata.
The AI coding agent then worked with the collected information to organize projects, determine which content belonged on which page, and implement the website.
Collect Instagram posts to understand the creator's work
I started with the Instagram latest posts template. Rather than manually opening every post and copying captions, the template returned structured information about the account's recent publications.
The available records included post links, content types, captions, image references, and other post information.
The captions were the most useful part. One explained the background of a short film, including how the creator became involved and the process behind producing it.
Another provided context for a voice-over project.
Several posts documented different stages of an event, from preparation to the final experience.
Once I had those captions, the coding agent could distinguish between individual posts and complete projects.
For example, instead of creating three unrelated cards for an event, I could group the content into a case study that explained the preparation, the event itself, and the resulting creative work.
That approach gave visitors a better understanding of what the creator contributed.
Use Instagram comments to add authentic audience feedback
A portfolio usually needs to show more than completed work. Potential clients may also want to know how people responded to that work.
For this project, I used Spidra's Instagram post comments template to retrieve available public comments from selected posts.
The collection returned 236 comment records across 23 posts from the two accounts.
These records included comment text, usernames, dates, and available engagement information.
That gave me a collection of actual audience reactions to review.
Some viewers commented on the creativity and production quality of the films. Others responded to the emotional impact of the work. Wedding posts included reactions to the storytelling and editing.
I used selected reactions in two ways.
First, I placed short comments near relevant projects to provide context from viewers.
Second, I built a guestbook-style section where visitors could see some of the feedback the creator had received.
The important distinction is that these were real comments, not AI-generated testimonials.
However, collecting a comment doesn't automatically make it suitable for publication. I still needed to review the text, ensure it was represented accurately, and consider whether it was appropriate to display it on a separate website.
Public comments are also not necessarily endorsements. When displaying them, it is better to preserve their original context rather than rewriting them into promotional statements.
This approach can be useful for portfolios, agency websites, event pages, and other projects where authentic feedback helps explain the response to published work.
Retrieve individual Instagram posts for richer content
The Templates provided an overview of the creator's work, but some sections needed more detailed information.
For those, I used Spidra's Scrape tool with individual Instagram post and reel links.
A structured post response can include the caption, media information, publication date, engagement counts, audio details, and other available metadata.
For the selected reels in this project, the results also contained video URLs that could be used to retrieve the media.
That helped solve another problem with the initial website.
I didn't want the portfolio to be a collection of external Instagram embeds. The creator's work needed to fit naturally into the website's design.
Using media that the creator was authorized to republish, I incorporated selected reels and carousel slides directly into the site.
This made it possible to build custom video presentations and tap-through carousels that matched the visual style of the portfolio.
It also meant visitors could explore the work without being redirected away from the website for every video.
Give the AI coding agent verified content to work with
Collecting the data was only half the workflow. The next challenge was making it useful to the AI coding agent.
A collection of Instagram records can contain captions, identifiers, media links, comments, and engagement information. But those records aren't automatically organized into portfolio projects.
The agent needs to understand what each record represents and how it might fit into the website.
For this project, that meant combining the extracted information with the creator's CV, project notes, and existing assets.
I wanted the agent to identify related posts, suggest project categories, and use the source material to prepare the content for different pages.
One useful approach is to give a coding agent an instruction like this:
I'm building a portfolio website for a content creator.
I've collected structured Instagram posts, captions, comments, and media information using Spidra. I also have the creator's CV and existing portfolio assets.
Review the information and identify the creator's projects, services, and types of work.
Group posts that belong to the same project. Use the original captions to understand the work and prepare short project descriptions.
Preserve the source URL for every project and comment. Do not invent project details, client names, results, or testimonials.
Suggest a website structure based on the actual work represented in the data. Flag information that needs clarification before building the pages.
Only use media that is approved for publication.
This is an example instruction you can adapt, rather than the exact prompt used in my project.
The main idea is to give the agent source material and clear boundaries before asking it to produce the final website.
Without those boundaries, an agent may turn incomplete information into confident-sounding descriptions that the source material doesn't support.
Let the data inform the website structure
One of the most useful outcomes was that the portfolio's structure changed as more information became available.
At the beginning, the obvious approach was to create a homepage, an About page, a project gallery, and a contact page.
That would have worked, but it wouldn't have represented the full range of the creator's work.
After collecting the Instagram data, I could identify several distinct areas of activity.
Films deserved their own presentation. Event coverage worked better as connected stories. Wedding projects needed a dedicated section. Styling and brand content had different purposes.
The website's services page could also reflect actual work instead of a generic list of creative services.
This didn't mean every Instagram post became its own portfolio item.
Some posts were useful as supporting context. Others were strong enough to become featured projects. A few were better left out.
That editorial selection made the website more useful than a complete archive of Instagram posts.
A portfolio should help someone understand what a creator does and why they might want to work with them. It doesn't need to reproduce the entire social media feed.
Applying the same approach to other websites
Although this project involved a content creator, the workflow can support other kinds of websites.
- A photographer could collect information about shoots and use captions to provide context for a gallery.
- A video production studio could group related social posts into campaign case studies, combining finished videos with behind-the-scenes material.
- A marketing agency could organize campaign content, published creative work, and selected audience reactions into a portfolio of client projects.
- An event organizer could use posts from an event to build a recap page with photographs, videos, and a timeline of activities.
- A small business could use its own approved social content to identify products, services, and customer stories that deserve dedicated website pages.
The details change, but the underlying process is similar.
You collect information from the relevant sources, turn it into structured records, verify the content, and use it to build something more useful than a collection of links.
Spidra handles the data collection part. The website's content, layout, and final presentation depend on what you're building.
From social media content to a useful website
The biggest improvement in this project came once the information was available.
The creator's captions helped explain completed projects. Posts from a second account revealed another business offering. Comments provided real audience reactions. Detailed post information helped bring the creative work into the website itself.
That gave the AI coding agent better information to work with and made it easier to build a portfolio that reflected the creator's actual experience.
The same approach can be useful whenever content is spread across social media but needs to be organized for a more specific purpose.
If you're building a portfolio, content platform, or another website that depends on public web information, explore Spidra Templates and the Instagram scraping capabilities to see how they can support your workflow.