Edge Impulse API
The Impulse API from Edge Impulse — 13 operation(s) for impulse.
Operations 18
Each operation below carries the questions people ask an LLM about it and the instructions they give an agent to run it. Generated by API Evangelist overlay
Ask an LLM
“What input, processing and learning blocks make up my project's default impulse?”
“Can I fetch one specific impulse instead of the default one?”
Tell an agent
Show the default impulse for project {projectId}.
Get impulse {impulseId} in project {projectId}, enabled blocks only.
Ask an LLM
“How do I define the full pipeline of input, signal processing and learning blocks for an Edge Impulse project?”
“Can I replace the blocks of an existing impulse in one call?”
Tell an agent
Set the impulse in project {projectId} to input blocks {inputBlocks}, DSP blocks {dspBlocks} and learn blocks {learnBlocks}.
Configure impulse {impulseId} with blocks {inputBlocks}, {dspBlocks} and {learnBlocks}.
Ask an LLM
“How do I wipe the impulse design of a project and start over?”
“Does clearing an impulse also remove its processing and learning blocks?”
Tell an agent destructive · confirm first
Clear the default impulse in project {projectId}.
Delete impulse {impulseId} and all its blocks from project {projectId}.
Ask an LLM
“Can I add another blank impulse to a project to try a different experiment?”
“How do I get the ID of a fresh, empty impulse?”
Tell an agent
Create a new empty impulse in project {projectId} and give me its ID.
Add a blank impulse experiment to project {projectId}.
Ask an LLM
“How do I rename one of my impulse experiments?”
“Can I add tags to an impulse to keep experiments organized?”
Tell an agent
Rename impulse {impulseId} in project {projectId} to {name}.
Tag the default impulse of project {projectId} with {tags}.
GET
/api/{projectId}/impulse/all
Get an impulse including disabled blocks · Get impulse including disabled blocks
#
Ask an LLM
“Can I see the blocks I disabled in my impulse, not just the active ones?”
“Which view of an impulse shows every block regardless of whether it is turned on?”
Tell an agent
Show the impulse for project {projectId} including disabled blocks.
Get impulse {impulseId} with its disabled blocks included.
POST
/api/{projectId}/impulse/get-new-block-id
Get an unused block ID for an impulse · Get new block ID
#
Ask an LLM
“How do I avoid reusing a block ID when building an impulse by hand?”
“What ID should a new block in my impulse get?”
Tell an agent
Give me an unused block ID for project {projectId}.
Reserve a fresh block number I can use when adding a block to project {projectId}.
GET
/api/{projectId}/impulse/blocks
List blocks available for building an impulse · Get impulse blocks
#
Ask an LLM
“Which processing and learning blocks can I choose from when designing an impulse?”
“What block types does Edge Impulse offer for my project?”
Tell an agent
List all blocks I can add to an impulse in project {projectId}.
Show the catalog of available impulse blocks for project {projectId}.
POST
/api/{projectId}/verify-dsp-block/url
Verify a custom DSP block URL · Verify custom DSP block
#
Ask an LLM
“How can I check that my self-hosted custom DSP block is reachable and valid?”
“Will my custom processing block at a given URL work before I add it to the impulse?”
Tell an agent
Verify the custom DSP block at {url} for project {projectId}.
Check whether {url} is a valid custom DSP block.
Ask an LLM
“How many impulse experiments does my project have?”
“Can I get a plain list of every impulse without accuracy figures?”
Tell an agent
List all impulses in project {projectId}.
Show me every impulse experiment defined in project {projectId}.
GET
/api/{projectId}/impulses-detailed
Compare impulses with accuracy and performance · Get all impulses (incl. metrics)
#
Ask an LLM
“Which of my impulse experiments has the best accuracy?”
“Can I see latency and memory metrics for every impulse side by side?”
Tell an agent
List the impulses in project {projectId} with their accuracy and performance metrics.
Compare all impulse experiments in project {projectId} by accuracy.
GET
/api/{projectId}/download-impulses-detailed
Download impulse metrics as JSON or CSV · Download all impulses (incl. metrics), as JSON or CSV
#
Ask an LLM
“Can I export my impulse experiment results to a CSV for a spreadsheet?”
“What file formats are available for downloading impulse metrics?”
Tell an agent
Download all impulses with metrics for project {projectId} as {format}.
Export the impulse comparison table of project {projectId} to a file.
POST
/api/{projectId}/impulse/legacy-set-impulse
Overwrite impulse state in the legacy format · Set legacy impulse
#
Ask an LLM
“Is there an internal way to write a project's impulse in the old pre-experiments format?”
“Which internal call clears all of a project's files and restores a legacy impulse from a zip?”
Tell an agent destructive · confirm first
Set the legacy impulse state of project {projectId} from zip {zip}, impulse {impulse} and config {config}.
Write impulse {impulse} to project {projectId} in the old single-impulse format with config {config} and zip {zip}.
POST
/api/{projectId}/impulse/{impulseId}/clone/structure
Clone an impulse's structure without data · Clone impulse (structure)
#
Ask an LLM
“Can I copy an impulse's blocks and config to a new experiment without copying the generated features?”
“How do I duplicate just the design of an impulse?”
Tell an agent
Clone the structure of impulse {impulseId} in project {projectId} as {name}.
Copy only the blocks and config of impulse {impulseId} into a new impulse called {name}.
POST
/api/{projectId}/impulse/{impulseId}/clone/complete
Clone an impulse including config and data · Clone impulse (complete)
#
Ask an LLM
“How do I make a full copy of an impulse, data and all?”
“Can I duplicate a trained impulse so I can tweak it without losing the original?”
Tell an agent
Fully clone impulse {impulseId} in project {projectId}, data included, as {name}.
Make a complete copy of impulse {impulseId} with its data and name it {name}.
GET
/api/{projectId}/transfer-learning-models
List available transfer learning models · Get all transfer learning models
#
Ask an LLM
“Which pretrained transfer learning models can I start from?”
“What transfer learning base models are offered across categories like image and audio?”
Tell an agent
List all transfer learning models available to project {projectId}.
Show the pretrained base models I can use for transfer learning in project {projectId}.
POST
/api/{projectId}/impulse/{impulseId}/set-thresholds
Set confidence and anomaly thresholds · Set thresholds
#
Ask an LLM
“How do I change the minimum confidence rating my model needs before it reports a class?”
“Can I raise the anomaly score threshold for an impulse and have model testing rerun automatically?”
Tell an agent
Set thresholds {thresholds} on impulse {impulseId} in project {projectId}, allowing regenerate jobs: {allowCreatingRegenerateModelTestingJobs}.
Update the min confidence and anomaly thresholds of impulse {impulseId} to {thresholds}.
POST
/api/{projectId}/impulse/{impulseId}/regenerate-model-testing-summary
Regenerate model testing results · Regenerate model testing summary
#
Ask an LLM
“I changed thresholds but no testing job started — how do I refresh the model testing results?”
“Can I recompute model testing without regenerating features?”
Tell an agent
Regenerate the model testing summary for impulse {impulseId} in project {projectId}.
Refresh test results of impulse {impulseId} after my threshold change.