Key Concepts
Learn the core Springbase concepts: Contexts, retrieval, Recipes, Plans, Agents, Canvas, Meetings, permissions, and reusable assets.
The Springbase Mental Model
Springbase works best when you understand three layers:
- Memory: the knowledge Springbase can reference.
- Execution: the ways Springbase can perform work.
- Output: the artifacts, automations, and shared assets that come out of that work.
Most users start with Chat. Power users build a system where Contexts, Recipes, Plans, Agents, Meetings, and Canvas reinforce each other.
Memory -> Execution -> Output -> Reuse
Memory Layer: Contexts
Contexts are how you teach Springbase what matters. Without Context, AI answers are generic. With Context, answers can be grounded in your actual docs, meetings, and connected tools.
| Concept | What it is | Why it matters |
|---|---|---|
| Contexts | Focused knowledge bases made from files, meeting transcripts, live sources, or community knowledge. | Keeps topics separated so Springbase can retrieve the right information for the right job. |
| Projects | Workspaces for organizing source material, documents, and collaboration around a goal. | Helps teams keep related knowledge and outputs together. |
| RAG / retrieval | The process Springbase uses to find relevant source chunks before answering. | Makes answers more specific and reduces hallucination. |
| Citations | Links back to the source material used in an answer. | Lets you verify claims instead of trusting the model blindly. |
| Live Sources | Connections to apps like Google Drive, Notion, GitHub, Slack, and other knowledge systems. | Keeps AI memory fresh when source material changes. |
| Community Contexts | Public Contexts shared by other Springbase users. | Lets you equip useful knowledge without building everything from scratch. |
Context Hygiene
Good Contexts are focused, named clearly, and kept clean.
- Use one Context per domain, team, client, or workflow.
- Remove outdated docs when they conflict with newer docs.
- Prefer source files with clear titles and useful structure.
- Use live sources when the knowledge changes frequently.
Execution Layer: Workflows
Once Springbase has memory, you choose the right execution surface for the job.
| Execution mode | How it works | Best used for |
|---|---|---|
| Chat | A flexible conversation with models, files, Contexts, and tools. | Exploration, brainstorming, fast questions, and one-off tasks. |
| Recipes | Reusable prompt templates with variables and optional Contexts. | Repeatable work like summaries, briefs, reports, reviews, and outbound copy. |
| Plan Mode | Springbase drafts an editable step-by-step plan before executing a complex goal. | Larger projects, multi-source research, content systems, strategy, and analysis. |
| Agents | AI workflows that can use connected tools to take action in other apps. | Sending updates, moving data, creating records, or coordinating across apps. |
| Scheduled Agents | Agents that run on a recurring schedule. | Monitoring sources, recurring reports, inbox cleanup, and operational routines. |
Choosing The Right Surface
| If you need to... | Use |
|---|---|
| Ask a quick question | Chat |
| Repeat the same prompt often | Recipe |
| Complete a multi-step project | Plan Mode |
| Take action in another app | Agent |
| Run something every day or week | Scheduled Agent |
| Edit and polish a substantial output | Canvas |
Workspace Layer: Canvas And Meetings
Springbase is not only a place to ask questions. It is also a place to capture real-world work and turn outputs into useful artifacts.
| Component | Function | Why it matters |
|---|---|---|
| Canvas | A workspace for documents, code, tables, diagrams, and HTML. | Important outputs can be edited, shared, exported, and reused instead of getting buried in chat. |
| Meetings | Transcription, summaries, decisions, action items, and post-meeting automation. | Turns conversations into searchable knowledge and follow-up work. |
| Solve | A lightweight public problem-solving surface. | Lets people experience Springbase-style structured problem solving quickly. |
| Model access | Access to many frontier and specialty models. | You can choose the right model for the task instead of being locked into one provider. |
Trust Layer: Permissions
Because Springbase can hold sensitive business knowledge, visibility and permissions matter.
| Visibility | Who can access it | Best for |
|---|---|---|
| Private | Only you. | Personal work, sensitive docs, private drafts, financial data. |
| Shared | Specific people or team members. | Collaborative team workflows and internal knowledge bases. |
| Unlisted | Anyone with the direct link. | Client reviews, contractor access, lightweight sharing. |
| Public | Discoverable by the community. | Helpful public Contexts, reusable Recipes, and showcase assets. |
Use the narrowest access level that still supports the workflow.
Marketplace And Reuse
Springbase becomes more powerful when good work is reusable.
| Asset | Reuse pattern |
|---|---|
| Context | Equip a knowledge base so it can ground Chat, Recipes, Plans, or Agents. |
| Recipe | Re-run a reliable prompt with different inputs. |
| Plan | Save a multi-step workflow that can be adapted later. |
| Canvas artifact | Share, export, or refine a finished output. |
Public assets can help other users move faster. Private assets can help your own team build internal operating leverage.
Models, Credits, And Cost Awareness
Springbase gives access to many models because different jobs need different strengths.
- Use fast, efficient models for drafts and everyday work.
- Use stronger reasoning models for strategy, analysis, coding, and complex planning.
- Use specialized image, audio, or video models when the output is not just text.
- Watch credit usage when running long Plans, large Context retrieval, or media generation.
The practical rule: use the lightest model that gets the job done, then upgrade when quality or reasoning matters.
The Most Important Terms
| Term | Plain-English meaning |
|---|---|
| Context | A focused memory bank for Springbase. |
| Source | A file, meeting, app connection, or webpage that feeds a Context. |
| Citation | A reference showing where an answer came from. |
| Recipe | A reusable AI workflow built from a prompt template. |
| Variable | A fill-in field inside a Recipe. |
| Plan | A multi-step execution path generated from a goal. |
| Agent | AI that can use tools and connected apps. |
| Canvas | A workspace for substantial outputs. |
| Artifact | A document, code file, table, diagram, or other created output. |
| Equip | Add a shared or community asset to your own workspace. |
A Simple First System
If you want to feel how the pieces connect:
- Create a Context called
Company Wiki. - Upload three useful docs.
- Ask Chat a question with
@CompanyWiki. - Save the best prompt as a Recipe.
- Use Plan Mode for a bigger goal based on the same Context.
- Send the final output to Canvas.
That is the basic Springbase loop: memory becomes execution, execution becomes output, and output becomes reusable leverage.
Related docs
Understand the Springbase ecosystem: Contexts, Chat, Recipes, Plans, Agents, Canvas, and the compounding loop that ties them together.
Get from first chat to a grounded Springbase workflow in minutes.
Learn how to build living knowledge bases in Springbase with documents, notes, shared access, live sources, community Contexts, and grounded Chat.
Turn reliable prompts into reusable templates with variables and optional Contexts.
Use Agents when workflows need connected app actions.