Back office ops · Production

Grab evolves SpellVault from no-code LLM app builder to agentic AI platform with MCP support

The problem

SpellVault's original no-code LLM app builder, running on a legacy single-step executor, was insufficient to keep pace with the rapidly evolving agentic AI landscape and the more complex, dynamic use cases teams needed to build.

Workflow diagram · grounded in source
1
No-code app creation
trigger
“empower everyone at Grab to effortlessly build and manage AI-powered apps without the need for coding”
2
RAG knowledge retrieval
ai_action
“SpellVault has built-in integrations with knowledge sources such as Wikis, Google Docs, as well as plain text and PDF uploads. These capabilities empower users to build assistants that reference relevant knowledge and provide more accura…”
3
Plugin data fetching
integration
“SpellVault plugins—modular components that allow apps to interact with internal systems (e.g. service dashboards, incident trackers) and external APIs (e.g. search engines, weather data)”
4
ReAct agent execution
ai_action
“This transformed all existing SpellVault apps into 'Reasoning and Acting' agents, better known as ReAct agents - a "one size fits many" solution that significantly enhanced the capabilities of these apps”
5
Deep Research search
ai_action
“The Deep Research capability came with SpellVault's ability to search across internal information repositories (e.g., Slack messages, Wiki, Jira) within Grab, as well as searching online for relevant information.”
6
MCP tool exposure
output
“Each app created in SpellVault can now be exposed through the MCP protocol. This allows other agents or MCP-compatible clients, such as IDEs or external orchestration frameworks, to treat a SpellVault app as a callable tool.”
Reported outcome

SpellVault evolved into a full agentic platform—with ReAct agents, a unified Native Tools framework, and MCP service support—enabling thousands of apps for automation, experimentation, and production use cases across Grab.

Reported metrics
apps created on SpellVaultthousands of apps
Reported stack
SpellVaultRAGSlackKibanaGoogle DocsPythonReAct agentsMCPTinyMCPFastAPIJiraDeep Research
Source
https://engineering.grab.com/spellvault-evolution-beyond-llm
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

SpellVault evolved into a full agentic platform—with ReAct agents, a unified Native Tools framework, and MCP service support—enabling thousands of apps for automation, experimentation, and production use cases across…

What tools did this team use?

SpellVault, RAG, Slack, Kibana, Google Docs, Python, ReAct agents, MCP, TinyMCP, FastAPI.

What results were reported?

apps created on SpellVault: thousands of apps (source-reported, not independently verified).

How is this back office ops AI workflow structured?

No-code app creation → RAG knowledge retrieval → Plugin data fetching → ReAct agent execution → Deep Research search → MCP tool exposure.

Related back office ops cases